Suggest executable actions in response to detected events

Abstract

The present disclosure relates to suggesting executable actions in response to detecting an event. Systems and processes for providing suggested user actions via an electronic device are disclosed. The suggested actions are provided in response to detecting a predefined event occurring in a user's day. An occurrence of an anchor is encoded in a signal generated by the electronic device. The occurrence of the anchor can be detected via monitoring and analyzing an electronic signal. Based on a user's previous interaction with the device, the occurrence of the anchor indicates user behavior and / or an action taken in response to the anchor. Machine learning (ML) is used to train an anchor model to associate actions taken in response to an anchor occurrence. The trained anchor model is used to detect anchors and provide suggested actions in response to the detected anchor occurrence. The suggested actions are based on the type of anchor occurrence and the contextual conditions of the anchor occurrence.

Description

Technical Field

[0001] The present invention generally relates to intelligent automated assistants, and more particularly to enabling a digital assistant to intelligently suggest executable actions in response to detecting events indicative of user behavior based on prior user interactions with an electronic device implementing the digital assistant. Background Art

[0002] Intelligent automated assistants (or digital assistants) can provide a favorable interface between human users and electronic devices. Such assistants can allow users to interact with devices or systems using natural language in voice form and / or text form. For example, a user can provide a voice input containing a user request to a digital assistant running on an electronic device. The digital assistant can interpret the user's intent from the voice input and operationalize the user's intent into tasks. These tasks can then be performed by executing one or more services of the electronic device, and relevant output in response to the user's request can be returned to the user. Summary of the invention

[0003] Example methods are disclosed herein. An example method includes training a prediction model (e.g., an anchor model) for a detected event (e.g., an anchor), the detected event indicating a user behavior of an electronic device capable of performing each action type in an enabled action type set. The electronic device may have one or more processors and a memory. The electronic device may perform one or more actions and / or operations. The method may include multiple training actions and multiple event occurrences of events detected at an electronic device and based on multiple signals generated by the electronic device. Each event in the multiple event occurrences may be associated with metadata indicating contextual conditions of the event occurrence. Each training action in the multiple training actions may be initiated via user interaction with the electronic device, and each training action in the multiple training actions may be classified as an action type in the enabled action type set. A candidate action type set may be determined. Determining the candidate action type set may be based on multiple correlations between each training action in the multiple training actions and each event occurrence in the multiple event occurrences. The candidate action type set may be a subset of the enabled action type set. The ranking of each action type in the candidate action type set may be determined. Determining the ranking of action types may be based on a portion of the multiple training actions classified as action types. Determine the order of action types can be further based on the one or more contextual conditions indicated by the metadata associated with a portion of the multiple events, and a portion of the multiple events is related to the portion of the multiple training actions via the multiple correlations. In some embodiments, the first action type of the candidate action type set can be selected based on the order determined for each action type of the candidate action type set. The first part of the multiple training actions and the first part of the multiple events can be selected. Each part in the first part of the multiple training actions can be classified as a first action type. Each part in the first part of the multiple events can be related to at least one of the first parts of the multiple training actions via the multiple correlations. The time offset of the first action type can be determined based on the time distribution of the first part of the multiple events relative to the first part of the multiple events. In some embodiments, the prediction model can be updated to generate a suggested action in response to another occurrence of the event. The suggested action provided is based on the time offset of the first action type and the first action type.

[0004] An example non-transitory computer-readable medium is disclosed herein. An example non-transitory computer-readable storage medium stores one or more programs. The one or more programs include instructions that, when executed by one or more processors of an electronic device, cause the electronic device to perform an action for training a prediction model for an event indicating user behavior of an electronic device capable of performing each action type in an enabled action type set. These actions may include multiple training actions and multiple event occurrences at an electronic device and based on multiple signals generated by the electronic device to detect an event. Each event in the multiple event occurrences may be associated with metadata indicating contextual conditions for the event occurrence. Each of the multiple training actions may be initiated via user interaction with an electronic device, and each of the multiple training actions may be classified as an action type in the enabled action type set. A candidate action type set may be determined. Determining the candidate action type set may be based on multiple correlations between each of the multiple training actions and each of the multiple event occurrences. The candidate action type set may be a subset of the enabled action type set. The ranking of each action type in the candidate action type set may be determined. Determining the ranking of action types may be based on a portion of the multiple training actions classified as action types. Determine the order of action types can be further based on the one or more contextual conditions indicated by the metadata associated with a portion of the multiple events, and a portion of the multiple events is related to the portion of the multiple training actions via the multiple correlations. In some embodiments, the first action type of the candidate action type set can be selected based on the order determined for each action type of the candidate action type set. The first part of the multiple training actions and the first part of the multiple events can be selected. Each part in the first part of the multiple training actions can be classified as a first action type. Each part in the first part of the multiple events can be related to at least one of the first parts of the multiple training actions via the multiple correlations. The time offset of the first action type can be determined based on the time distribution of the first part of the multiple events relative to the first part of the multiple events. In some embodiments, the prediction model can be updated to generate a suggested action in response to another occurrence of the event. The suggested action provided is based on the time offset of the first action type and the first action type.

[0005] An example electronic device is disclosed herein. An example electronic device includes one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for performing operations for training a prediction model for events indicating user behavior of an electronic device capable of performing each action type in an enabled action type set. These operations may include multiple training actions and multiple event occurrences at an electronic device and based on multiple signals generated by the electronic device to detect events. Each event in the multiple event occurrences may be associated with metadata indicating contextual conditions of the event occurrence. Each of the multiple training actions may be initiated via user interaction with the electronic device, and each of the multiple training actions may be classified as an action type in the enabled action type set. A candidate action type set may be determined. Determining the candidate action type set may be based on multiple correlations between each of the multiple training actions and each of the multiple event occurrences. The candidate action type set may be a subset of the enabled action type set. The ranking of each action type in the candidate action type set may be determined. Determining the ranking of action types may be based on a portion of the multiple training actions classified as action types. Determine the order of action types can be further based on the one or more contextual conditions indicated by the metadata associated with a portion of the multiple events, and a portion of the multiple events is related to the portion of the multiple training actions via the multiple correlations. In some embodiments, the first action type of the candidate action type set can be selected based on the order determined for each action type of the candidate action type set. The first part of the multiple training actions and the first part of the multiple events can be selected. Each part in the first part of the multiple training actions can be classified as a first action type. Each part in the first part of the multiple events can be related to at least one of the first parts of the multiple training actions via the multiple correlations. The time offset of the first action type can be determined based on the time distribution of the first part of the multiple events relative to the first part of the multiple events. In some embodiments, the prediction model can be updated to generate a suggested action in response to another occurrence of the event. The suggested action provided is based on the time offset of the first action type and the first action type.

[0006] Another example method may be used to employ a predictive model for an event indicative of user behavior of an electronic device. Other methods may include detecting an event occurrence of the event based on one or more signals generated by the electronic device. The event occurrence may be associated with metadata indicating contextual conditions of the event occurrence. Based on the event occurrence, a suggested action and a time offset may be received from the predictive model. The suggested action may be provided to the user within a time offset relative to the event occurrence.

[0007] Another example non-transitory computer-readable storage medium stores one or more other programs. The one or more other programs include instructions that, when executed by one or more processors of an electronic device, cause the electronic device to perform other operations for employing a predictive model for an event indicating user behavior of the electronic device. Other operations may include detecting an event occurrence of the event based on one or more signals generated by the electronic device. The event occurrence may be associated with metadata indicating contextual conditions of the event occurrence. Based on the event occurrence, a suggested action and a time offset may be received from the predictive model. The suggested action may be provided to the user within a time offset relative to the event occurrence.

[0008] Another example electronic device includes one or more processors; a memory; and one or more other programs, wherein the one or more other programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include instructions for performing other operations for adopting a predictive model for an event indicating user behavior of the electronic device. Other operations may include detecting an event occurrence of the event based on one or more signals generated by the electronic device. The event occurrence may be associated with metadata indicating contextual conditions of the event occurrence. Based on the event occurrence, a suggested action and a time offset may be received from the predictive model. The suggested action may be provided to the user within a time offset relative to the event occurrence. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A block diagram is shown of a system and environment for implementing a digital assistant according to various examples.

[0010] Figure 2A A block diagram of a portable multifunction device that implements the client-side portion of a digital assistant according to various examples is shown.

[0011] Figure 2B is a block diagram illustrating exemplary components for event processing according to various examples.

[0012] Figure 3 A portable multifunction device implementing the client-side portion of a digital assistant according to various examples is shown.

[0013] Figure 4 is a block diagram of an exemplary multifunction device with a display and a touch-sensitive surface according to various examples.

[0014] Figure 5A An exemplary user interface of a menu of applications on a portable multifunction device is shown according to various examples.

[0015] Figure 5BAn exemplary user interface for a multifunction device having a touch-sensitive surface separate from the display is shown according to various examples.

[0016] Fig. 6A A personal electronic device according to various examples is shown.

[0017] Figure 6B A block diagram illustrating a personal electronic device according to various examples is shown.

[0018] Fig. 7A A block diagram of a digital assistant system or a server portion thereof according to various examples is shown.

[0019] Figure 7B According to various examples, Fig. 7A The capabilities of the digital assistant shown in .

[0020] Figure 7C A portion of an ontology according to various examples is shown.

[0021] Fig. 8A A classification decision tree is shown for the action type of playing a specific podcast and the anchor type of the user waking up according to various examples.

[0022] Figure 8B A classification decision tree of the action type of playing a specific song and the anchor type of the user waking up according to various examples is shown.

[0023] Fig. 9 A bar graph showing the temporal distribution of action types of playing a specific podcast in conjunction with wakeup anchors according to various examples.

[0024] Fig.10 A process for training a predictive model that enables providing suggested actions in response to the occurrence of an anchor event is shown according to various examples.

[0025] Fig.11 A process 1100 for deploying a trained predictive model at an electronic device is shown according to various examples. DETAILED DESCRIPTION

[0026] In the following description of the examples, reference will be made to the accompanying drawings, in which specific examples that may be implemented are shown by way of illustration. It should be understood that other examples may be used and structural changes may be made without departing from the scope of the various examples.

[0027] Embodiments relate to providing suggested user actions via electronic devices. Suggested actions may be provided in response to detecting the occurrence of an anchor event. An anchor event (or simply anchor) may be a predefined event that occurs during a user's day. For example, upon waking up in the morning (e.g., user waking up may be an example of an anchor), an electronic device may provide a notification including a suggestion for playing the latest episode of a particular podcast (i.e., a suggested user action), and the user often listens to a particular podcast when waking up in the morning. The occurrence of an anchor (or event) may be detected via one or more signals generated by the electronic device (e.g., the alarm function of the device is being executed, the device is converted from "do not disturb mode" to "active mode", etc.). The occurrence of an anchor may be detected via monitoring and analyzing electronic signals generated by the electronic device. Based on the user's previous interaction with the device (e.g., encoded in training data), the occurrence of an anchor may indicate user behavior and / or actions taken in response to events during the user's day. As a non-limiting example, an anchor may include a user input location of interest (LOI), such as, but not limited to, a user's home, office, fitness center, airport, shopping mall, etc. Because an anchor may be an event that indicates user behavior, an anchor may be referred to as an anchor event or simply an event throughout the text. The occurrence of an anchor may be referred to as an anchor occurrence.

[0028] Other non-limiting examples of the occurrence of an anchor include a user starting to use an electronic device after an extended period of device idle time, a user ending a workout, a user waking up, a user going to bed, a user pairing a Bluetooth-enabled device with an electronic device, a user starting or completing a calendar event, and the like. Other examples of anchors include a user launching a specific application of a device and / or adopting a specific function of a device. Any event that can be detected via a signature encoded in a signal generated by an electronic device can be an anchor. The signature indicating the occurrence of a specific anchor (e.g., a pattern within an electronic signal generated by a device) can be predetermined, pre-calculated, and / or pre-learned (e.g., via supervised or unsupervised machine learning (ML) methods). Therefore, an electronic device can monitor its various signals and detect the occurrence of a specific anchor by identifying and / or detecting one or more signatures within a signal indicating a specific anchor. For example, the occurrence of a wakeup anchor can be detected via a signal of an electronic device indicating that a user has terminated or otherwise ended the "idle" state of the electronic device (or the alarm of the device rings). Upon detecting the occurrence of an anchor, various embodiments can suggest (via previously generated training data) one or more user actions associated with the occurrence of the anchor.

[0029] A user action (i.e., an action) may include invoking, executing, or otherwise initiating a specific application, capability, function, or command enabled via an electronic device. As non-limiting examples, a user action may include, but is not limited to, playing specific audio / video content (e.g., a podcast, a music playlist, an audiobook, a concert, a television series, a movie, etc.), launching an application installed on a device (e.g., a fitness application, a meditation application, a ride-sharing application, a food delivery application, a social networking application, etc.), sending an electronic message (e.g., an email, SMS, Twitter, etc.) to another user, a user group, or a social network, turning on / off a function of a device (e.g., turning off / on airplane mode of a device), updating various settings and / or configurations of a device, creating a calendar event, etc.

[0030] When providing a suggested action, various embodiments may provide a notification (e.g., a "pop-up" notification) indicating the suggested action. The notification may be interactive in that the user may initiate the action via interactive selection of the notification. Upon selecting the notification, the suggested action may be performed by the device and / or execution of the action may be initiated.

[0031] In various embodiments, one or more machine learning (ML) models may be trained to learn various statistical correlations (or associations) between the occurrence of a particular anchor and user-initiated actions (e.g., innovation of device functionality or capabilities). Any of the various ML models employed by various embodiments may be collectively referred to herein as "anchor models" because these models are trained to detect the occurrence of various anchors and suggest one or more actions that are statistically likely to be initiated by the user in response to the occurrence (or imminent occurrence) of the anchor. Therefore, various embodiments include the training of anchor models, and the use of trained anchor models to enhance the user's experience (UX) of using the device. Supervised or unsupervised ML may be used to train the model to determine the context (e.g., contextual features) in which the anchor occurs, and to establish, customize, or selectively target suggested actions based on the context of the anchor.

[0032] Although the following description uses the terms "first," "second," etc. to describe various elements, these elements should not be limited by the terms. These terms are simply used to distinguish one element from another. For example, without departing from the scope of the various described examples, a first input may be referred to as a second input, and similarly, a second input may be referred to as a first input. Both the first input and the second input are inputs, and in some cases are independent and different inputs.

[0033] The terms used in the description of various examples described herein are only for the purpose of describing specific examples and are not intended to be limiting. As used in the description of various examples described and in the appended claims, the singular forms "a", "an" and "the" are intended to also include plural forms, unless the context clearly indicates otherwise. It will also be understood that the term "and / or" used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items. It will also be understood that the term "includes", "including", "comprises" and / or "comprising" when used in this specification specifies the presence of stated features, integers, steps, operations, elements and / or parts, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, parts and / or their groupings.

[0034] Depending on the context, the term "if" may be interpreted to mean "when" or "upon" or "in response to determining" or "in response to detecting." Similarly, the phrases "if it is determined that ..." or "if [a stated condition or event] is detected" may be interpreted to mean "upon determining that ..." or "in response to determining that ..." or "upon detecting [a stated condition or event]" or "in response to detecting [a stated condition or event]," depending on the context.

[0035] 1. System and Environment

[0036] Figure 1 A block diagram of a system 100 according to various examples is shown. In some examples, the system 100 implements a digital assistant. The terms "digital assistant," "virtual assistant," "intelligent automated assistant," or "automatic digital assistant" refer to any information processing system that interprets natural language input in spoken and / or textual form to infer user intent and performs actions based on the inferred user intent. For example, to act on an inferred user intent, the system performs one or more of the following steps: identifying a task flow with steps and parameters designed to implement the inferred user intent, inputting specific requirements into the task flow according to the inferred user intent; executing the task flow by calling a program, method, service, API, etc.; and generating an output response to the user in an audible (e.g., voice) and / or visual form.

[0037] Specifically, the digital assistant is capable of accepting user requests in the form of at least partially natural language commands, requests, statements, narrations and / or inquiries. Typically, the user requests to seek an informative answer or perform a task from the digital assistant. A satisfactory response to the user's request includes providing the requested informative answer, performing the requested task, or a combination of the two. For example, the user asks the digital assistant a question, such as "Where am I now?" Based on the user's current location, the digital assistant answers "You are near the West Gate of Central Park." The user also requests to perform a task, such as "Please invite my friends to my girlfriend's birthday party next week." In response, the digital assistant can confirm the request by saying "OK, right away", and then send a suitable calendar invitation to each of the user's friends listed in the user's electronic address book on behalf of the user. During the execution of the requested task, the digital assistant sometimes interacts with the user in a continuous dialogue involving multiple information exchanges over a long period of time. There are many other methods of interacting with the digital assistant to request information or perform various tasks. In addition to providing verbal responses and taking programmed actions, the digital assistant also provides responses in other video or audio forms, such as text, alarms, music, videos, animations, etc.

[0038] like Figure 1 As shown, in some examples, the digital assistant is implemented according to a client-server model. The digital assistant includes a client-side portion 102 (hereinafter referred to as "DA client 102") executed on a user device 104 and a server-side portion 106 (hereinafter referred to as "DA server 106") executed on a server system 108. The DA client 102 communicates with the DA server 106 via one or more networks 110. The DA client 102 provides client-side functionality, such as user-oriented input and output processing, and communication with the DA server 106. The DA server 106 provides server-side functionality for any number of DA clients 102, each located on a corresponding user device 104.

[0039] In some examples, the DA server 106 includes a client-facing I / O interface 112, one or more processing modules 114, data and models 116, and an I / O interface 118 to external services. The client-facing I / O interface 112 facilitates client-facing input and output processing of the DA server 106. One or more processing modules 114 utilize the data and models 116 to process voice input and determine user intent based on natural language input. In addition, one or more processing modules 114 perform task execution based on inferred user intent. In some examples, the DA server 106 communicates with external services 120 via one or more networks 110 to complete tasks or collect information. The I / O interface 118 to external services facilitates such communications.

[0040] User device 104 can be any suitable electronic device. In some examples, user device 104 is a portable multi-function device (e.g., Figure 2A The device 200 described above), a multifunctional device (eg, as described below with reference to Figure 4 The device 400) or a personal electronic device (eg, as described below) FIG. 6A to FIG. 6B The portable multifunction device is, for example, a mobile phone that also includes other functions such as a PDA and / or a music player. Specific examples of portable multifunction devices include the Apple iPod and Device. Other examples of portable multifunction devices include, but are not limited to, earbuds / headphones, speakers, and laptops or tablets. In addition, in some examples, user device 104 is a non-portable multifunction device. Specifically, user device 104 is a desktop computer, a game console, a speaker, a television, or a television set-top box. In some examples, user device 104 includes a touch-sensitive surface (e.g., a touch screen display and / or a trackpad). In addition, user device 104 optionally includes one or more other physical user interface devices, such as a physical keyboard, a mouse, and / or a joystick. Various examples of electronic devices such as multifunction devices are described in more detail below.

[0041] Examples of one or more communication networks 110 include local area networks (LANs) and wide area networks (WANs), such as the Internet. One or more communication networks 110 are implemented using any known network protocols, including various wired or wireless protocols, such as Ethernet, Universal Serial Bus (USB), FIREWIRE, Global System for Mobile Communications (GSM), Enhanced Data GSM Environment (EDGE), Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Bluetooth, Wi-Fi, Voice over Internet Protocol (VoIP), Wi-MAX, or any other suitable communication protocol.

[0042] The server system 108 is implemented on one or more stand-alone data processing devices or distributed computer networks. In some examples, the server system 108 also uses various virtual devices and / or services of third-party service providers (e.g., third-party cloud service providers) to provide the potential computing resources and / or infrastructure resources of the server system 108.

[0043] In some examples, the user device 104 communicates with the DA server 106 via a second user device 122. The second user device 122 is similar or identical to the user device 104. For example, the second user device 122 is similar to the user device 104 described below with reference to Figure 2A , Figure 4 and FIG. 6A to FIG. 6B The device 200, 400 or 600 described. The user device 104 is configured to be communicatively coupled to the second user device 122 via a direct communication connection (such as Bluetooth, NFC, BTLE, etc.) or via a wired or wireless network (such as a local Wi-Fi network). In some examples, the second user device 122 is configured to act as a proxy between the user device 104 and the DA server 106. For example, the DA client 102 of the user device 104 is configured to transmit information (e.g., a user request received at the user device 104) to the DA server 106 via the second user device 122. The DA server 106 processes the information and returns relevant data (e.g., data content in response to the user request) to the user device 104 via the second user device 122.

[0044] In some examples, the user device 104 is configured to send an abbreviated request for data to a second user device 122 to reduce the amount of information transmitted from the user device 104. The second user device 122 is configured to determine supplemental information to add to the abbreviated request to generate a complete request for transmission to the DA server 106. The system architecture can advantageously allow user devices 104 with limited communication capabilities and / or limited battery power (e.g., a watch or similar compact electronic device) to access services provided by the DA server 106 by using a second user device 122 (e.g., a mobile phone, laptop, tablet, etc.) with greater communication capabilities and / or battery power as a proxy to the DA server 106. Although Figure 1 Only two user devices 104 and 122 are shown, but it should be understood that in some examples, the system 100 may include any number and type of user devices configured to communicate with the DA server system 106 in this proxy configuration.

[0045] Although Figure 1 The digital assistant shown in includes both a client-side portion (e.g., DA client 102) and a server-side portion (e.g., DA server 106), but in some examples, the functionality of the digital assistant is implemented as a stand-alone application installed on a user device. In addition, the division of functionality between the client and server portions of the digital assistant may vary in different specific implementations. For example, in some examples, the DA client is a thin client that only provides user-facing input and output processing functions and delegates all other functions of the digital assistant to a backend server.

[0046] 2. Electronic devices

[0047] Attention now turns to embodiments of electronic devices for implementing the client-side portion of a digital assistant. Figure 2A is a block diagram showing a portable multifunction device 200 with a touch-sensitive display system 212 according to some embodiments. The touch-sensitive display 212 is sometimes referred to as a "touch screen" for convenience, and is sometimes referred to as or referred to as a "touch-sensitive display system." The device 200 includes a memory 202 (which optionally includes one or more computer-readable storage media), a memory controller 222, one or more processing units (CPUs) 220, a peripheral device interface 218, an RF circuit 208, an audio circuit 210, a speaker 211, a microphone 213, an input / output (I / O) subsystem 206, other input control devices 216, and an external port 224. The device 200 optionally includes one or more optical sensors 264. The device 200 optionally includes one or more contact force sensors 265 for detecting the intensity of contact on the device 200 (e.g., a touch-sensitive surface of the device 200 such as the touch-sensitive display system 212). Device 200 optionally includes one or more tactile output generators 267 for generating tactile output on device 200 (e.g., generating tactile output on a touch-sensitive surface such as touch-sensitive display system 212 of device 200 or touch pad 455 of device 400). These components optionally communicate via one or more communication buses or signal lines 203.

[0048] As used in this specification and claims, the term "intensity" of a contact on a touch-sensitive surface refers to the force or pressure (force per unit area) of a contact (e.g., a finger contact) on the touch-sensitive surface, or to a surrogate (surrogate) of the force or pressure of a contact on the touch-sensitive surface. The intensity of a contact has a range of values ​​that includes at least four different values ​​and more typically includes hundreds of different values ​​(e.g., at least 256). The intensity of a contact is optionally determined (or measured) using various methods and various sensors or combinations of sensors. For example, one or more force sensors below or adjacent to the touch-sensitive surface are optionally used to measure the force at different points on the touch-sensitive surface. In some implementations, force measurements from multiple force sensors are combined (e.g., weighted averaged) to determine an estimated contact force. Similarly, the pressure-sensitive tip of a stylus is optionally used to determine the pressure of the stylus on the touch-sensitive surface. Alternatively, the size of the contact area detected on the touch-sensitive surface and / or its change, the capacitance of the touch-sensitive surface near the contact and / or its change, and / or the resistance of the touch-sensitive surface near the contact and / or its change are optionally used as a substitute for the force or pressure of the contact on the touch-sensitive surface. In some embodiments, the substitute measurement of the contact force or pressure is used directly to determine whether the intensity threshold has been exceeded (e.g., the intensity threshold is described in units corresponding to the substitute measurement). In some embodiments, the substitute measurement of the contact force or pressure is converted into an estimated force or pressure, and the estimated force or pressure is used to determine whether the intensity threshold has been exceeded (e.g., the intensity threshold is a pressure threshold measured in units of pressure). Using the intensity of the contact as an attribute of the user input allows the user to access additional device functions that would otherwise be inaccessible to the user on a smaller device with limited real estate, which is used to display an enable indication (e.g., on a touch-sensitive display) and / or receive user input (e.g., via a touch-sensitive display, a touch-sensitive surface, or a physical control / mechanical control, such as a knob or button).

[0049] As used in this specification and claims, the term "tactile output" refers to a physical displacement of a device relative to a previous position of the device, a physical displacement of a component of the device (e.g., a touch-sensitive surface) relative to another component of the device (e.g., a housing), or a displacement of a component relative to the center of mass of the device, which will be detected by a user using the user's sense of touch. For example, in the case where a device or a component of the device is in contact with a user's touch-sensitive surface (e.g., a finger, palm, or other part of the user's hand), the tactile output generated by the physical displacement will be interpreted by the user as a tactile sensation corresponding to a perceived change in a physical characteristic of the device or component of the device. For example, movement of a touch-sensitive surface (e.g., a touch-sensitive display or trackpad) is optionally interpreted by the user as a "press click" or "release click" to a physical actuation button. In some cases, the user will feel a tactile sensation, such as a "press click" or "release click", even when the physical actuation button associated with the touch-sensitive surface that is physically pressed (e.g., displaced) by the user's movement does not move. As another example, even when there is no change in the smoothness of the touch-sensitive surface, movement of the touch-sensitive surface may optionally be interpreted or sensed by the user as "roughness" of the touch-sensitive surface. Although such a user's interpretation of touch will be limited by the user's individualized sensory perceptions, many sensory perceptions of touch are common to most users. Thus, when a tactile output is described as corresponding to a particular sensory perception of a user (e.g., "press click," "release click," "roughness"), unless otherwise stated, the generated tactile output corresponds to a physical displacement of the device or a component thereof that would generate that sensory perception for a typical (or average) user.

[0050] It should be understood that device 200 is only one example of a portable multifunction device, and that device 200 optionally has more or fewer components than shown, optionally combines two or more components, or optionally has a different configuration or arrangement of the components. Figure 2A The various components shown in the EMBODIMENTS 2000 are implemented in hardware, software, or a combination of hardware and software, including one or more signal processing and / or application specific integrated circuits.

[0051] The memory 202 includes one or more computer-readable storage media. These computer-readable storage media are, for example, tangible and non-transitory. The memory 202 includes a high-speed random access memory and also includes a non-volatile memory, such as one or more disk storage devices, flash memory devices, or other non-volatile solid-state memory devices. The memory controller 222 controls other components of the device 200 to access the memory 202.

[0052] In some examples, the non-transitory computer-readable storage medium of memory 202 is used to store instructions (e.g., for performing various aspects of the processes described below) for use by or in conjunction with an instruction execution system, device, or apparatus, such as a computer-based system, a system containing a processor, or other system that can fetch instructions from an instruction execution system, device, or apparatus and execute the instructions. In other examples, the instructions (e.g., for performing various aspects of the processes described below) are stored on a non-transitory computer-readable storage medium (not shown) of server system 108, or are divided between the non-transitory computer-readable storage medium of memory 202 and the non-transitory computer-readable storage medium of server system 108.

[0053] The peripheral device interface 218 is used to couple the input and output peripheral devices of the device to the CPU 220 and the memory 202. The one or more processors 220 run or execute various software programs and / or instruction sets stored in the memory 202 to perform various functions of the device 200 and process data. In some embodiments, the peripheral device interface 218, the CPU 220, and the memory controller 222 are implemented on a single chip such as the chip 204. In some other embodiments, they are implemented on separate chips.

[0054] RF (radio frequency) circuit 208 receives and sends RF signals, also known as electromagnetic signals. RF circuit 208 converts electrical signals into / converts electromagnetic signals into electrical signals, and communicates with communication networks and other communication devices via electromagnetic signals. RF circuit 208 optionally includes well-known circuits for performing these functions, including but not limited to antenna systems, RF transceivers, one or more amplifiers, tuners, one or more oscillators, digital signal processors, codec chipsets, user identity modules (SIM) cards, memory, etc. RF circuit 208 optionally communicates with networks and other devices through wireless communications, such as the Internet (also known as the World Wide Web (WWW)), intranets, and / or wireless networks (such as cellular telephone networks, wireless local area networks (LANs), and / or metropolitan area networks (MANs)). RF circuit 208 optionally includes well-known circuits for detecting near field communication (NFC) fields, such as through short-range communication radio components. Wireless communication optionally uses any of a variety of communication standards, protocols, and technologies, including, but not limited to, Global System for Mobile Communications (GSM), Enhanced Data GSM Environment (EDGE), High Speed ​​Downlink Packet Access (HSDPA), High Speed ​​Uplink Packet Access (HSUPA), Evolution, Data-Only (EV-DO), HSPA, HSPA+, Dual Cell HSPA (DC-HSPDA), Long Term Evolution (LTE), Near Field Communication (NFC), Wideband Code Division Multiple Access (W-CDMA), Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Bluetooth, Bluetooth Low Energy (BTLE), Wireless Fidelity (Wi-Fi) (e.g., IEEE 802.11a, IEEE 802.11b, IEEE 802.11g, IEEE 802.11d), IEEE 802.11e, IEEE 802.11f, IEEE 802.11g, IEEE 802.11g). 802.11n and / or IEEE 802.11ac), Voice over Internet Protocol (VoIP), Wi-MAX, email protocols (e.g., Internet Message Access Protocol (IMAP) and / or Post Office Protocol (POP)), instant messaging (e.g., Extensible Messaging and Presence Protocol (XMPP), Session Initiation Protocol for Instant Messaging and Presence Utilizing Extensions (SIMPLE), Instant Messaging and Presence Service (IMPS)) and / or Short Message Service (SMS), or any other suitable communication protocol, including communication protocols that have not yet been developed as of the filing date of this document.

[0055] The audio circuit 210, speaker 211, and microphone 213 provide an audio interface between a user and the device 200. The audio circuit 210 receives audio data from the peripheral device interface 218, converts the audio data into electrical signals, and transmits the electrical signals to the speaker 211. The speaker 211 converts the electrical signals into sound waves audible to humans. The audio circuit 210 also receives electrical signals converted from sound waves by the microphone 213. The audio circuit 210 converts the electrical signals into audio data and transmits the audio data to the peripheral device interface 218 for processing. The audio data is retrieved from and / or transmitted to the memory 202 and / or the RF circuit 208 via the peripheral device interface 218. In some embodiments, the audio circuit 210 also includes a headset jack (e.g., Figure 3 The headset jack provides an interface between the audio circuit 210 and a removable audio input / output peripheral device, such as an output-only headset or a headset with both output (e.g., a single-ear headset or a dual-ear headset) and input (e.g., a microphone).

[0056] The I / O subsystem 206 couples input / output peripherals on the device 200, such as a touch screen 212 and other input control devices 216, to a peripheral device interface 218. The I / O subsystem 206 optionally includes a display controller 256, an optical sensor controller 258, an intensity sensor controller 259, a tactile feedback controller 261, and one or more input controllers 260 for other input or control devices. One or more input controllers 260 receive / send electrical signals from other input control devices 216. Other input control devices 216 optionally include physical buttons (e.g., push buttons, rocker buttons, etc.), dials, slide switches, joysticks, click wheels, etc. In some alternative embodiments, the input controller 260 is optionally coupled to any of the following (or none of the following): a keyboard, an infrared port, a USB port, and a pointing device such as a mouse. One or more buttons (e.g., Figure 3 308 in ) optionally includes an up / down button for volume control of the speaker 211 and / or the microphone 213. The one or more buttons optionally include a push button (e.g., Figure 3 306 in ).

[0057] A quick press of the push button may disengage the lock of the touch screen 212 or begin the process of using gestures on the touch screen to unlock the device, as described in U.S. Patent Application No. 11 / 322,549, filed on December 23, 2005, entitled "Unlocking a Device by Performing Gestures on an Unlock Image," which is hereby incorporated by reference in its entirety. A longer press of the push button (e.g., 306) turns the device 200 on or off. The user can customize the function of one or more buttons. The touch screen 212 is used to implement virtual buttons or soft buttons and one or more soft keyboards.

[0058] The touch-sensitive display 212 provides an input interface and an output interface between the device and the user. The display controller 256 receives electrical signals from the touch screen 212 and / or sends electrical signals to the touch screen 212. The touch screen 212 displays visual output to the user. The visual output includes graphics, text, icons, videos, and any combination thereof (collectively referred to as "graphics"). In some embodiments, some or all of the visual output corresponds to user interface objects.

[0059] The touch screen 212 has a touch-sensitive surface, sensor, or sensor group that accepts input from a user based on tactile and / or haptic contact. The touch screen 212 and display controller 256 (together with any associated modules and / or instruction sets in memory 202) detect contact (and any movement or interruption of the contact) on the touch screen 212 and convert the detected contact into interaction with a user interface object (e.g., one or more soft keys, icons, web pages, or images) displayed on the touch screen 212. In an exemplary embodiment, the point of contact between the touch screen 212 and the user corresponds to the user's finger.

[0060] The touch screen 212 uses LCD (liquid crystal display) technology, LPD (light emitting polymer display) technology, or LED (light emitting diode) technology, although other display technologies may be used in other embodiments. The touch screen 212 and display controller 256 use any of a variety of touch sensing technologies now known or later developed, as well as other proximity sensor arrays or other elements for determining one or more points of contact with the touch screen 212 to detect contact and any movement or interruption thereof, including but not limited to capacitive, resistive, infrared, and surface acoustic wave technologies. In an exemplary embodiment, projected mutual capacitance sensing technology is used, such as in the Apple ® from Apple Inc. (Cupertino, California). and iPod The technology used in.

[0061] In some embodiments, the touch-sensitive display of the touch screen 212 is similar to the multi-touch-sensitive trackpads described in U.S. Patents 6,323,846 (Westerman et al.), 6,570,557 (Westerman et al.), and / or 6,677,932 (Westerman), and / or U.S. Patent Publication 2002 / 0015024A1, all of which are hereby incorporated by reference in their entirety. However, the touch screen 212 displays visual output from the device 200, whereas the touch-sensitive trackpad does not provide visual output.

[0062] In some embodiments, the touch-sensitive display of the touch screen 212 is as described in the following applications: (1) U.S. patent application No. 11 / 381,313, filed on May 2, 2006, entitled "Multipoint Touch Surface Controller"; (2) U.S. patent application No. 10 / 840,862, filed on May 6, 2004, entitled "Multipoint Touchscreen"; (3) U.S. patent application No. 10 / 903,964, filed on July 30, 2004, entitled "Gestures For Touch Sensitive Input Devices"; (4) U.S. patent application No. 11 / 048,264, filed on January 31, 2005, entitled "Gestures For Touch Sensitive Input Devices"; (5) U.S. patent application No. 11 / 051,964, filed on January 18, 2005, entitled "Mode-Based Graphical User Interfaces For Touch Sensitive Input Devices"; and (9) U.S. patent application No. 11 / 367,749, filed on March 3, 2006, entitled “Multi-Functional Hand-Held Device.” All of these applications are hereby incorporated by reference in their entirety.

[0063] The touch screen 212 has, for example, a video resolution of more than 100 dpi. In some embodiments, the touch screen has a video resolution of about 160 dpi. The user contacts the touch screen 212 using any suitable object or appendage, such as a stylus, a finger, etc. In some embodiments, the user interface is designed to work primarily through finger-based contacts and gestures, which may not be as precise as stylus-based input due to the larger contact area of ​​a finger on the touch screen. In some embodiments, the device converts rough finger-based input into precise pointer / cursor positions or commands for performing the actions desired by the user.

[0064] In some embodiments, in addition to the touch screen, the device 200 also includes a touch pad (not shown) for activating or deactivating specific functions. In some embodiments, the touch pad is a touch-sensitive area of ​​the device that, unlike the touch screen, does not display visual output. The touch pad is a touch-sensitive surface that is separate from the touch screen 212 or an extension of the touch-sensitive surface formed by the touch screen.

[0065] Device 200 also includes a power system 262 for powering the various components. Power system 262 includes a power management system, one or more power sources (e.g., batteries, alternating current (AC)), a recharging system, power fault detection circuits, power converters or inverters, power status indicators (e.g., light emitting diodes (LEDs)), and any other components associated with the generation, management, and distribution of power in a portable device.

[0066] The device 200 also includes one or more optical sensors 264 . Figure 2A An optical sensor coupled to an optical sensor controller 258 in the I / O subsystem 206 is shown. The optical sensor 264 includes a charge coupled device (CCD) or a complementary metal oxide semiconductor (CMOS) phototransistor. The optical sensor 264 receives light projected through one or more lenses from the environment and converts the light into data representing an image. In conjunction with the imaging module 243 (also called a camera module), the optical sensor 264 captures a static image or video. In some embodiments, the optical sensor is located at the rear of the device 200, opposite to the touch screen display 212 at the front of the device, so that the touch screen display is used as a viewfinder for static images and / or video image acquisition. In some embodiments, the optical sensor is located at the front of the device so that the user's image is obtained for video conferencing while the user views other video conference participants on the touch screen display. In some embodiments, the position of the optical sensor 264 can be changed by the user (e.g., by rotating the lens and sensor in the device housing) so that a single optical sensor 264 is used with the touch screen display for both video conferencing and static image and / or video image acquisition.

[0067] Device 200 optionally also includes one or more contact intensity sensors 265 . Figure 2A A contact force sensor is shown coupled to a force sensor controller 259 in the I / O subsystem 206. The contact force sensor 265 optionally includes one or more piezoresistive strain gauges, capacitive force sensors, electrical force sensors, piezoelectric force sensors, optical force sensors, capacitive touch-sensitive surfaces, or other force sensors (e.g., sensors for measuring the force (or pressure) of contact on a touch-sensitive surface). The contact force sensor 265 receives contact force information (e.g., pressure information or a surrogate for pressure information) from the environment. In some embodiments, at least one contact force sensor is juxtaposed or adjacent to a touch-sensitive surface (e.g., touch-sensitive display system 212). In some embodiments, at least one contact force sensor is located on the rear of the device 200, opposite the touch screen display 212 located on the front of the device 200.

[0068] Device 200 also includes one or more proximity sensors 266 . Figure 2A A proximity sensor 266 is shown coupled to the peripherals interface 218. Alternatively, the proximity sensor 266 is coupled to the input controller 260 in the I / O subsystem 206. The proximity sensor 266 is implemented as described in the following U.S. patent applications: 11 / 241,839, entitled "Proximity Detector In Handheld Device"; 11 / 240,788, entitled "Proximity Detector In Handheld Device"; 11 / 620,702, entitled "Using Ambient Light Sensor To Augment Proximity Sensor Output"; 11 / 586,862, entitled "Automated Response To And Sensing Of User Activity In Portable Devices"; and 11 / 638,251, entitled "Methods And Systems For Automatic Configuration Of Peripherals", which are hereby incorporated by reference in their entirety. In some embodiments, when the multifunction device is placed near the user's ear (e.g., when the user is making a phone call), the proximity sensor turns off and the touch screen 212 is disabled.

[0069] Device 200 optionally also includes one or more tactile output generators 267. Figure 2AA tactile output generator coupled to a tactile feedback controller 261 in the I / O subsystem 206 is shown. The tactile output generator 267 optionally includes one or more electroacoustic devices such as a speaker or other audio component; and / or an electromechanical device for converting energy into linear motion such as a motor, solenoid, electroactive polymer, piezoelectric actuator, electrostatic actuator, or other tactile output generating component (e.g., a component for converting an electrical signal into a tactile output on the device). The contact force sensor 265 receives tactile feedback generation instructions from the tactile feedback module 233 and generates a tactile output on the device 200 that can be felt by a user of the device 200. In some embodiments, at least one tactile output generator is arranged in juxtaposition or proximity to a touch-sensitive surface (e.g., a touch-sensitive display system 212) and optionally generates a tactile output by moving the touch-sensitive surface vertically (e.g., inward / outward of the surface of the device 200) or laterally (e.g., backward and forward in the same plane as the surface of the device 200). In some embodiments, at least one tactile output generator sensor is located on the back of the device 200, opposite the touch screen display 212 located on the front of the device 200.

[0070] Device 200 also includes one or more accelerometers 268 . Figure 2A An accelerometer 268 coupled to the peripheral device interface 218 is shown. Alternatively, the accelerometer 268 is coupled to the input controller 260 in the I / O subsystem 206. The accelerometer 268 performs as described in the following U.S. patent publications: U.S. Patent Publication No. 20050190059, "Acceleration-based Theft Detection System for Portable Electronic Devices" and U.S. Patent Publication No. 20060017692, "Methods And Apparatuses For Operating A Portable Device Based On An Accelerometer", both of which are incorporated herein by reference in their entirety. In some embodiments, information is displayed in a portrait view or a landscape view on the touch screen display based on analysis of data received from one or more accelerometers. The device 200 optionally includes a magnetometer (not shown) and a GPS (or GLONASS or other global navigation system) receiver (not shown) in addition to one or more accelerometers 268 for obtaining information about the position and orientation (e.g., portrait or landscape) of the device 200.

[0071] In some embodiments, the software components stored in memory 202 include an operating system 226, a communication module (or instruction set) 228, a contact / motion module (or instruction set) 230, a graphics module (or instruction set) 232, a text input module (or instruction set) 234, a global positioning system (GPS) module (or instruction set) 235, a digital assistant client module 229, and an application (or instruction set) 236. In addition, memory 202 stores data and models, such as user data and models 231. In addition, in some embodiments, memory 202 ( Figure 2A ) or 470( Figure 4 ) storage device / global internal state 257, such as Figure 2A and Figure 4 . The device / global internal state 257 includes one or more of the following: an active application state, which indicates which applications (if any) are currently active; a display state, which indicates what applications, views, or other information occupy various areas of the touch screen display 212; a sensor state, including information obtained from the device's various sensors and input control devices 216; and position information about the device's position and / or posture.

[0072] The operating system 226 (e.g., Darwin, RTXC, LINUX, UNIX, OS X, iOS, WINDOWS, or an embedded operating system such as VxWorks) includes various software components and / or drivers for controlling and managing general system tasks (e.g., memory management, storage device control, power management, etc.), and facilitating communication between various hardware components and software components.

[0073] The communication module 228 facilitates communication with other devices through one or more external ports 224, and also includes various software components for processing data received by the RF circuit 208 and / or the external port 224. The external port 224 (e.g., Universal Serial Bus (USB), FireWire, etc.) is suitable for coupling directly to other devices, or indirectly through a network (e.g., the Internet, wireless LAN, etc.). In some embodiments, the external port is connected to (trademark of Apple Inc.) devices.

[0074] The contact / motion module 230 optionally detects contact with the touch screen 212 (in conjunction with the display controller 256) and other touch-sensitive devices (e.g., a touchpad or physical click wheel). The contact / motion module 230 includes various software components for performing various operations related to contact detection, such as determining whether contact has occurred (e.g., detecting a finger press event), determining the contact strength (e.g., the force or pressure of the contact, or a substitute for the force or pressure of the contact), determining whether there is movement of the contact and tracking the movement on the touch-sensitive surface (e.g., detecting one or more finger drag events), and determining whether the contact has stopped (e.g., detecting a finger lift event or contact disconnection). The contact / motion module 230 receives contact data from the touch-sensitive surface. Determining the movement of the contact point optionally includes determining the rate (magnitude), velocity (magnitude and direction), and / or acceleration (change in magnitude and / or direction) of the contact point, the movement of the contact point being represented by a series of contact data. These operations are optionally applied to a single point of contact (e.g., a single finger contact) or multiple points of simultaneous contact (e.g., "multi-touch" / multiple finger contacts). In some embodiments, the contact / motion module 230 and display controller 256 detect contact on the touch pad.

[0075] In some embodiments, the contact / motion module 230 uses a set of one or more intensity thresholds to determine whether an operation has been performed by a user (e.g., to determine whether a user has "clicked" an icon). In some embodiments, at least a subset of the intensity thresholds are determined based on software parameters (e.g., the intensity thresholds are not determined by the activation thresholds of specific physical actuators and can be adjusted without changing the physical hardware of the device 200). For example, without changing the touchpad or touchscreen display hardware, the mouse "click" threshold of a touchpad or touchscreen can be set to any one of a large range of predefined thresholds. In addition, in some specific implementations, a software setting is provided to the user of the device for adjusting one or more intensity thresholds in a set of intensity thresholds (e.g., by adjusting individual intensity thresholds and / or by utilizing a system-level click on an "intensity" parameter to adjust multiple intensity thresholds at once).

[0076] The contact / motion module 230 optionally detects gesture input from a user. Different gestures on a touch-sensitive surface have different contact patterns (e.g., different motions, timings, and / or intensities of the detected contacts). Therefore, gestures are optionally detected by detecting specific contact patterns. For example, detecting a finger tap gesture includes detecting a finger press event, and then detecting a finger lift (lift-off) event at the same location (or substantially the same location) as the finger press event (e.g., at the location of an icon). As another example, detecting a finger swipe gesture on a touch-sensitive surface includes detecting a finger press event, then detecting one or more finger drag events, and then detecting a finger lift (lift-off) event.

[0077] The graphics module 232 includes various known software components for rendering and displaying graphics on the touch screen 212 or other display, including components for changing the visual impact (e.g., brightness, transparency, saturation, contrast, or other visual characteristics) of the displayed graphics. As used herein, the term "graphics" includes any object that can be displayed to a user, including, without limitation, text, web pages, icons (such as user interface objects including soft keys), digital images, videos, animations, etc.

[0078] In some embodiments, the graphics module 232 stores data representing graphics to be used. Each graphic is optionally assigned a corresponding code. The graphics module 232 receives one or more codes specifying the graphics to be displayed from an application program, etc., along with coordinate data and other graphic attribute data if necessary, and then generates screen image data to be output to the display controller 256.

[0079] The tactile feedback module 233 includes various software components for generating instructions used by the one or more tactile output generators 267 to produce tactile outputs at one or more locations on the device 200 in response to user interaction with the device 200 .

[0080] The text input module 234, which in some examples is a component of the graphics module 232, provides a soft keyboard for entering text in various applications (e.g., contacts 237, email 240, IM 241, browser 247, and any other application requiring text input).

[0081] The GPS module 235 determines the location of the device and provides this information for use in various applications (e.g., to the phone 238 for use in location-based dialing; to the camera 243 as picture / video metadata; and to applications that provide location-based services, such as the Weather widget, the Local Pages widget, and the Maps / Navigation widget).

[0082] The digital assistant client module 229 includes various client-side digital assistant instructions to provide the client-side functions of the digital assistant. For example, the digital assistant client module 229 can accept sound input (e.g., voice input), text input, touch input and / or gesture input through various user interfaces of the portable multifunction device 200 (e.g., microphone 213, one or more accelerometers 268, touch-sensitive display system 212, one or more optical sensors 264, other input control devices 216, etc.). The digital assistant client module 229 can also provide output in the form of audio (e.g., voice output), output in the form of visual output and / or output in the form of tactile output through various output interfaces of the portable multifunction device 200 (e.g., speaker 211, touch-sensitive display system 212, one or more tactile output generators 267, etc.). For example, the output is provided as voice, sound, alarm, text message, menu, graphic, video, animation, vibration and / or a combination of two or more of the above. During operation, the digital assistant client module 229 uses the RF circuit 208 to communicate with the DA server 106.

[0083] User data and models 231 include various data associated with the user (e.g., user-specific vocabulary data, user preference data, user-specified name pronunciations, data from the user's electronic address book, to-do items, shopping lists, etc.) to provide client-side functionality of the digital assistant. In addition, user data and models 231 include various models for processing user input and determining user intent (e.g., speech recognition models, statistical language models, natural language processing models, knowledge ontologies, task flow models, service models, etc.).

[0084] In some examples, the digital assistant client module 229 utilizes the various sensors, subsystems, and peripherals of the portable multifunction device 200 to collect additional information from the surrounding environment of the portable multifunction device 200 to establish a context associated with the user, the current user interaction, and / or the current user input. In some examples, the digital assistant client module 229 provides the context information, or a subset thereof, to the DA server 106 along with the user input to help infer the user's intent. In some examples, the digital assistant also uses the context information to determine how to prepare and transmit output to the user. The context information is referred to as context data.

[0085] In some examples, contextual information accompanying the user input includes sensor information, such as lighting, ambient noise, ambient temperature, images or videos of the surrounding environment, etc. In some examples, the contextual information may also include the physical state of the device, such as device orientation, device location, device temperature, power level, speed, acceleration, motion pattern, cellular signal strength, etc. In some examples, information related to the software state of the DA server 106, such as the running process of the portable multifunction device 200, installed programs, past and current network activities, background services, error logs, resource usage, etc., is provided to the DA server 106 as contextual information associated with the user input.

[0086] In some examples, the digital assistant client module 229 selectively provides information stored on the portable multifunction device 200 (e.g., user data 231) in response to a request from the DA server 106. In some examples, the digital assistant client module 229 also elicits additional input from the user via a natural language dialog or other user interface upon request by the DA server 106. The digital assistant client module 229 transmits the additional input to the DA server 106 to assist the DA server 106 in inferring intent and / or implementing the user intent expressed in the user request.

[0087] Reference below 7A to 7C A more detailed description of the digital assistant is provided. It should be appreciated that the digital assistant client module 229 may include any number of sub-modules of the digital assistant module 726 described below.

[0088] The application 236 includes the following modules (or instruction sets) or a subset or superset thereof:

[0089] Contacts module 237 (sometimes referred to as address book or contact list);

[0090] Telephone module 238;

[0091] Video conferencing module 239;

[0092] Email client module 240;

[0093] Instant messaging (IM) module 241;

[0094] Fitness support module 242;

[0095] A camera module 243 for still images and / or video images;

[0096] Image management module 244;

[0097] Video player module;

[0098] Music player module;

[0099] Browser module 247;

[0100] Calendar module 248;

[0101] A widget module 249, which in some examples includes one or more of the following: a weather widget 249-1, a stock widget 249-2, a calculator widget 249-3, an alarm clock widget 249-4, a dictionary widget 249-5, and other widgets acquired by the user and a widget 249-6 created by the user;

[0102] A widget creator module 250 for forming a user-created widget 249-6;

[0103] Search module 251;

[0104] Video and music player module 252, which merges the video player module and the music player module;

[0105] Notepad module 253;

[0106] Map module 254; and / or

[0107] Online video module 255.

[0108] Examples of other applications 236 stored in memory 202 include other word processing applications, other image editing applications, drawing applications, rendering applications, JAVA-enabled applications, encryption, digital rights management, voice recognition, and voice reproduction.

[0109] In combination with the touch screen 212, display controller 256, touch / motion module 230, graphics module 232, and text input module 234, the contact module 237 is used to manage an address book or contact list (for example, stored in the application internal state 292 of the contact module 237 in memory 202 or memory 470), including: adding one or more names to the address book; deleting names from the address book; associating telephone numbers, email addresses, physical addresses or other information with names; associating images with names; categorizing and classifying names; providing telephone numbers or email addresses to initiate and / or facilitate communications via telephone 238, video conferencing module 239, email 240 or IM 241; and the like.

[0110] In conjunction with RF circuit 208, audio circuit 210, speaker 211, microphone 213, touch screen 212, display controller 256, contact / motion module 230, graphics module 232, and text input module 234, phone module 238 is used to enter a character sequence corresponding to a phone number, access one or more phone numbers in contact module 237, modify an already entered phone number, dial a corresponding phone number, conduct a conversation, and disconnect or hang up when the conversation is completed. As described above, wireless communications use any of a variety of communication standards, protocols, and technologies.

[0111] In combination with the RF circuit 208, the audio circuit 210, the speaker 211, the microphone 213, the touch screen 212, the display controller 256, the optical sensor 264, the optical sensor controller 258, the contact / motion module 230, the graphics module 232, the text input module 234, the contact module 237 and the telephone module 238, the video conferencing module 239 includes executable instructions for initiating, conducting and terminating a video conference between a user and one or more other participants in accordance with user instructions.

[0112] In conjunction with RF circuit 208, touch screen 212, display controller 256, contact / motion module 230, graphics module 232, and text input module 234, email client module 240 includes executable instructions for creating, sending, receiving, and managing emails in response to user instructions. In conjunction with image management module 244, email client module 240 makes it very easy to create and send emails with still images or video images captured by camera module 243.

[0113] In conjunction with the RF circuit 208, the touch screen 212, the display controller 256, the contact / motion module 230, the graphics module 232, and the text input module 234, the instant messaging module 241 includes executable instructions for the following operations: inputting a character sequence corresponding to an instant message, modifying previously input characters, transmitting a corresponding instant message (e.g., using a short message service (SMS) or multimedia message service (MMS) protocol for telephone-based instant messaging or using XMPP, SIMPLE, or IMPS for Internet-based instant messaging), receiving instant messages, and viewing received instant messages. In some embodiments, the transmitted and / or received instant messages include graphics, photos, audio files, video files, and / or other attachments such as those supported in MMS and / or enhanced messaging services (EMS). As used herein, "instant messaging" refers to both telephone-based messages (e.g., messages sent using SMS or MMS) and Internet-based messages (e.g., messages sent using XMPP, SIMPLE, or IMPS).

[0114] In combination with the RF circuit 208, the touch screen 212, the display controller 256, the contact / motion module 230, the graphics module 232, the text input module 234, the GPS module 235, the map module 254, and the music player module, the fitness support module 242 includes executable instructions for: creating a workout (e.g., with time, distance, and / or calorie burn goals); communicating with fitness sensors (sports equipment); receiving fitness sensor data; calibrating sensors for monitoring fitness; selecting and playing music for fitness; and displaying, storing, and transmitting fitness data.

[0115] In combination with the touch screen 212, the display controller 256, one or more optical sensors 264, the optical sensor controller 258, the touch / motion module 230, the graphics module 232, and the image management module 244, the camera module 243 includes executable instructions for the following operations: capturing still images or videos (including video streams) and storing them in the memory 202, modifying the characteristics of the still images or videos, or deleting the still images or videos from the memory 202.

[0116] In conjunction with the touch screen 212, display controller 256, touch / motion module 230, graphics module 232, text input module 234, and camera module 243, the image management module 244 includes executable instructions for arranging, modifying (e.g., editing), or otherwise manipulating, labeling, deleting, presenting (e.g., in a digital slide show or album), and storing still images and / or video images.

[0117] In combination with the RF circuit 208, the touch screen 212, the display controller 256, the touch / motion module 230, the graphics module 232 and the text input module 234, the browser module 247 includes executable instructions for browsing the Internet in accordance with user instructions, including searching, linking to, receiving and displaying web pages or portions thereof, as well as attachments and other files linked to web pages.

[0118] In combination with the RF circuit 208, the touch screen 212, the display controller 256, the touch / motion module 230, the graphics module 232, the text input module 234, the email client module 240 and the browser module 247, the calendar module 248 includes executable instructions for creating, displaying, modifying and storing calendars and data associated with the calendar (e.g., calendar entries, to-do items, etc.) in accordance with user instructions.

[0119] In combination with the RF circuit 208, the touch screen 212, the display controller 256, the contact / motion module 230, the graphics module 232, the text input module 234, and the browser module 247, the desktop widget module 249 is a mini-application that can be downloaded and used by the user (e.g., the weather desktop widget 249-1, the stock desktop widget 249-2, the calculator desktop widget 249-3, the alarm desktop widget 249-4, and the dictionary desktop widget 249-5) or a mini-application created by the user (e.g., the user-created desktop widget 249-6). In some embodiments, the desktop widget includes an HTML (Hypertext Markup Language) file, a CSS (Cascading Style Sheets) file, and a JavaScript file. In some embodiments, the desktop widget includes an XML (Extensible Markup Language) file and a JavaScript file (e.g., the Yahoo! desktop widget).

[0120] In combination with the RF circuit 208, the touch screen 212, the display controller 256, the touch / motion module 230, the graphics module 232, the text input module 234 and the browser module 247, the desktop widget creator module 250 is used by users to create desktop widgets (for example, to turn a user-specified portion of a web page into a desktop widget).

[0121] In combination with the touch screen 212, display controller 256, contact / motion module 230, graphics module 232 and text input module 234, the search module 251 includes executable instructions for searching the memory 202 for text, music, sound, images, videos and / or other files that match one or more search criteria (e.g., one or more user-specified search terms) in accordance with user instructions.

[0122] In conjunction with touch screen 212, display controller 256, contact / motion module 230, graphics module 232, audio circuitry 210, speaker 211, RF circuitry 208, and browser module 247, video and music player module 252 includes executable instructions that allow a user to download and play back recorded music and other sound files stored in one or more file formats (such as MP3 or AAC files), as well as executable instructions for displaying, presenting, or otherwise playing back video (e.g., on touch screen 212 or on an external display connected via external port 224). In some embodiments, device 200 optionally includes the functionality of an MP3 player such as an iPod (trademark of Apple Inc.).

[0123] In conjunction with the touch screen 212, display controller 256, contact / motion module 230, graphics module 232 and text input module 234, the notepad module 253 includes executable instructions for creating and managing notes, to-do lists, etc. according to user instructions.

[0124] In combination with the RF circuit 208, the touch screen 212, the display controller 256, the touch / motion module 230, the graphics module 232, the text input module 234, the GPS module 235 and the browser module 247, the map module 254 is used to receive, display, modify and store maps and data associated with the maps (e.g., driving directions, data related to stores and other points of interest at or near a particular location, and other location-based data) in accordance with user instructions.

[0125] In conjunction with the touch screen 212, display controller 256, contact / motion module 230, graphics module 232, audio circuit 210, speaker 211, RF circuit 208, text input module 234, email client module 240, and browser module 247, the online video module 255 includes instructions that allow a user to access, browse, receive (e.g., by streaming and / or downloading), play back (e.g., on the touch screen or on a connected external display via external port 224), send an email with a link to a specific online video, and otherwise manage online videos in one or more file formats (such as H.264). In some embodiments, the instant messaging module 241 is used instead of the email client module 240 to send a link to a specific online video. Additional descriptions of online video applications can be found in U.S. Provisional Patent Application No. 60 / 936,562, filed on June 20, 2007, entitled “Portable Multifunction Device, Method, and Graphical User Interface for Playing Online Videos,” and U.S. Patent Application No. 11 / 968,067, filed on December 31, 2007, entitled “Portable Multifunction Device, Method, and Graphical User Interface for Playing Online Videos,” the contents of which are hereby incorporated by reference in their entirety.

[0126] Each of the modules and applications described above corresponds to an executable instruction set for performing one or more of the functions described above and the methods described in this patent application (e.g., the computer-implemented methods and other information processing methods described herein). These modules (e.g., instruction sets) do not have to be implemented as separate software programs, processes, or modules, and therefore various subsets of these modules may be combined or otherwise rearranged in various embodiments. For example, a video player module may be combined with a music player module into a single module (e.g., Figure 2AIn some embodiments, the memory 202 stores a subset of the above modules and data structures. In addition, the memory 202 stores additional modules and data structures not described above.

[0127] In some embodiments, the device 200 is a device on which the operation of a predefined set of functions is performed exclusively through a touch screen and / or a touch pad. By using a touch screen and / or a touch pad as the primary input control device for the operation of the device 200, the number of physical input control devices (such as push buttons, dials, etc.) on the device 200 is reduced.

[0128] A predefined set of functions that are performed exclusively through a touch screen and / or a touchpad optionally includes navigation between user interfaces. In some embodiments, the touchpad, when touched by a user, navigates the device 200 to a main menu, a home menu, or a root menu from any user interface displayed on the device 200. In such embodiments, a touchpad is used to implement a "menu button." In some other embodiments, the menu button is a physical push button or other physical input control device, rather than a touchpad.

[0129] Figure 2B is a block diagram illustrating exemplary components for event processing according to some embodiments. In some embodiments, memory 202 ( Figure 2A ) or memory 470( Figure 4 ) includes an event classifier 270 (e.g., in the operating system 226) and a corresponding application 236-1 (e.g., any one of the aforementioned applications 237 to 251, 255, 480 to 490).

[0130] The event classifier 270 receives event information and determines the application 236-1 and the application view 291 of the application 236-1 to which the event information is to be delivered. The event classifier 270 includes an event monitor 271 and an event distributor module 274. In some embodiments, the application 236-1 includes an application internal state 292 that indicates one or more current application views displayed on the touch-sensitive display 212 when the application is active or executing. In some embodiments, the device / global internal state 257 is used by the event classifier 270 to determine which application(s) is currently active, and the application internal state 292 is used by the event classifier 270 to determine the application view 291 to which the event information is to be delivered.

[0131] In some embodiments, the application internal state 292 includes additional information, such as one or more of the following: resumption information to be used when the application 236-1 resumes execution, user interface state information indicating that information is being displayed or is ready to be displayed by the application 236-1, a state queue for enabling a user to return to a previous state or view of the application 236-1, and a repeat / undo queue of previous actions taken by the user.

[0132] Event monitor 271 receives event information from peripherals interface 218. Event information includes information about sub-events (e.g., a user touch on touch-sensitive display 212 as part of a multi-touch gesture). Peripherals interface 218 transmits information it receives from I / O subsystem 206 or sensors such as proximity sensor 266, one or more accelerometers 268, and / or microphone 213 (through audio circuit 210). The information that peripherals interface 218 receives from I / O subsystem 206 includes information from touch-sensitive display 212 or a touch-sensitive surface.

[0133] In some embodiments, event monitor 271 sends requests to peripheral device interface 218 at predetermined intervals. In response, peripheral device interface 218 transmits event information. In other embodiments, peripheral device interface 218 transmits event information only when there is a significant event (e.g., receiving an input above a predetermined noise threshold and / or receiving an input for more than a predetermined duration).

[0134] In some embodiments, the event classifier 270 also includes a hit view determination module 272 and / or an active event identifier determination module 273.

[0135] When the touch-sensitive display 212 displays more than one view, the hit view determination module 272 provides software procedures for determining where within one or more views a sub-event has occurred. A view consists of controls and other elements that a user can see on the display.

[0136] Another aspect of the user interface associated with an application is a set of views, sometimes referred to herein as application views or user interface windows, in which information is displayed and touch-based gestures occur. The application views (of the respective application) in which a touch is detected correspond to a programmatic level within a programmatic hierarchy of applications or a view hierarchy. For example, the lowest level view in which a touch is detected is called a hit view, and the set of events that are considered to be correct input is determined at least in part based on the hit view of the initial touch that started the touch-based gesture.

[0137] The hit view determination module 272 receives information related to sub-events of touch-based gestures. When an application has multiple views organized in a hierarchy, the hit view determination module 272 identifies the hit view as the lowest view in the hierarchy where the sub-events should be processed. In most cases, the hit view is the lowest level view in which the initiating sub-event (e.g., the first sub-event in a sequence of sub-events that form an event or potential event) occurs. Once a hit view is identified by the hit view determination module 272, the hit view typically receives all sub-events related to the same touch or input source for which it was identified as the hit view.

[0138] Active event recognizer determination module 273 determines which view or views within the view hierarchy should receive a particular sequence of sub-events. In some embodiments, active event recognizer determination module 273 determines that only the hit view should receive a particular sequence of sub-events. In other embodiments, active event recognizer determination module 273 determines that all views that include the physical location of the sub-event are actively participating views, and therefore determines that all actively participating views should receive a particular sequence of sub-events. In other embodiments, even if a touch sub-event is completely confined to an area associated with one particular view, higher views in the hierarchy will still remain as actively participating views.

[0139] Event distributor module 274 distributes event information to event identifiers (e.g., event identifier 280). In embodiments including active event identifier determination module 273, event distributor module 274 delivers the event information to the event identifier determined by active event identifier determination module 273. In some embodiments, event distributor module 274 stores the event information in an event queue, which is retrieved by corresponding event receiver 282.

[0140] In some embodiments, operating system 226 includes event classifier 270. Alternatively, application 236-1 includes event classifier 270. In yet another embodiment, event classifier 270 is a standalone module or is part of another module stored in memory 202, such as contact / motion module 230.

[0141] In some embodiments, the application 236-1 includes multiple event handlers 290 and one or more application views 291, each of which includes instructions for handling touch events occurring in a corresponding view of the user interface of the application. Each application view 291 of the application 236-1 includes one or more event identifiers 280. Typically, the corresponding application view 291 includes multiple event identifiers 280. In other embodiments, one or more event identifiers in the event identifier 280 are part of an independent module, which is a higher-level object such as a user interface toolkit (not shown) or the application 236-1 from which methods and other properties are inherited. In some embodiments, the corresponding event handler 290 includes one or more of the following: data updater 276, object updater 277, GUI updater 278, and / or event data 279 received from the event classifier 270. The event handler 290 utilizes or calls the data updater 276, object updater 277, or GUI updater 278 to update the application internal state 292. Alternatively, one or more of the application views in the application views 291 include one or more corresponding event handlers 290. In addition, in some embodiments, one or more of the data updater 276, the object updater 277, and the GUI updater 278 are included in the corresponding application view 291.

[0142] The corresponding event identifier 280 receives event information (e.g., event data 279) from the event classifier 270 and identifies the event from the event information. The event identifier 280 includes an event receiver 282 and an event comparator 284. In some embodiments, the event identifier 280 also includes metadata 283 and at least a subset of event delivery instructions 288 (which includes sub-event delivery instructions).

[0143] The event receiver 282 receives event information from the event classifier 270. The event information includes information about sub-events such as touches or touch movements. Depending on the sub-event, the event information also includes additional information, such as the location of the sub-event. When the sub-event involves the movement of a touch, the event information also includes the rate and direction of the sub-event. In some embodiments, the event includes the device rotating from one orientation to another orientation (e.g., from a longitudinal orientation to a transverse orientation, or vice versa), and the event information includes corresponding information about the current orientation of the device (also referred to as the device posture).

[0144] Event comparator 284 compares event information with predefined event or sub-event definitions, and based on the comparison, determines an event or sub-event, or determines or updates the state of an event or sub-event. In some embodiments, event comparator 284 includes event definition 286. Event definition 286 includes the definition of an event (e.g., a predefined sequence of sub-events), such as event 1 (287-1), event 2 (287-2), and other events. In some embodiments, sub-events in event (287) include, for example, touch start, touch end, touch move, touch cancel, and multi-touch. In one example, the definition of event 1 (287-1) is a double-click on a displayed object. For example, a double-click includes a first touch (touch start) of a predetermined duration on a displayed object, a first lift-off (touch end) of a predetermined duration, a second touch (touch start) of a predetermined duration on a displayed object, and a second lift-off (touch end) of a predetermined duration. In another example, the definition of event 2 (287-2) is a drag on a displayed object. For example, dragging includes a touch (or contact) of a predetermined duration on a displayed object, movement of the touch on the touch-sensitive display 212, and lifting of the touch (touch end). In some embodiments, the event also includes information for one or more associated event handlers 290.

[0145] In some embodiments, event definition 287 includes a definition of an event for a corresponding user interface object. In some embodiments, event comparator 284 performs a hit test to determine which user interface object is associated with a sub-event. For example, in an application view displaying three user interface objects on touch-sensitive display 212, when a touch is detected on touch-sensitive display 212, event comparator 284 performs a hit test to determine which of the three user interface objects is associated with the touch (sub-event). If each displayed object is associated with a corresponding event handler 290, the event comparator uses the result of the hit test to determine which event handler 290 should be activated. For example, event comparator 284 selects an event handler associated with a sub-event and the object that triggered the hit test.

[0146] In some embodiments, the definition of the corresponding event (287) also includes a delay action that delays the delivery of the event information until it has been determined that the sub-event sequence does or does not correspond to the event type of the event identifier.

[0147] When a corresponding event recognizer 280 determines that a sequence of sub-events does not match any event in event definition 286, the corresponding event recognizer 280 enters an event impossible, event failed, or event ended state, after which subsequent sub-events of the touch-based gesture are ignored. In this case, other event recognizers (if any) that remain active for the hit view continue to track and process sub-events of the ongoing touch-based gesture.

[0148] In some embodiments, the corresponding event identifier 280 includes metadata 283 with configurable properties, flags, and / or lists that indicate how the event delivery system should perform sub-event delivery to actively participating event identifiers. In some embodiments, metadata 283 includes configurable properties, flags, and / or lists that indicate how event identifiers interact or can interact with each other. In some embodiments, metadata 283 includes configurable properties, flags, and / or lists that indicate whether sub-events are delivered to different levels in a view or programmatic hierarchy.

[0149] In some embodiments, when one or more specific sub-events of an event are identified, the corresponding event identifier 280 activates the event handler 290 associated with the event. In some embodiments, the corresponding event identifier 280 delivers the event information associated with the event to the event handler 290. Activating the event handler 290 is different from sending (and deferred sending) the sub-events to the corresponding hit view. In some embodiments, the event identifier 280 throws a tag associated with the identified event, and the event handler 290 associated with the tag obtains the tag and performs a predefined process.

[0150] In some embodiments, event delivery instructions 288 include sub-event delivery instructions that deliver event information about a sub-event without activating an event handler. Instead, the sub-event delivery instruction delivers the event information to an event handler associated with a sub-event sequence or to an actively participating view. An event handler associated with a sub-event sequence or with an actively participating view receives the event information and executes a predetermined process.

[0151] In some embodiments, data updater 276 creates and updates data used in application 236-1. For example, data updater 276 updates phone numbers used in contact module 237 or stores video files used in video player module. In some embodiments, object updater 277 creates and updates objects used in application 236-1. For example, object updater 277 creates new user interface objects or updates the position of user interface objects. GUI updater 278 updates the GUI. For example, GUI updater 278 prepares display information and sends the display information to graphics module 232 for display on a touch-sensitive display.

[0152] In some embodiments, event handler 290 includes or has access to data updater 276, object updater 277, and GUI updater 278. In some embodiments, data updater 276, object updater 277, and GUI updater 278 are included in a single module of the corresponding application 236-1 or application view 291. In other embodiments, they are included in two or more software modules.

[0153] It should be understood that the above discussion of event processing of user touches on a touch-sensitive display also applies to other forms of user input that utilize input devices to operate the multifunction device 200, and not all user input is initiated on the touch screen. For example, mouse movement and mouse button presses, optionally in conjunction with single or multiple keyboard presses or hold; contact movement on a touch pad, such as tapping, dragging, scrolling, etc.; stylus input; movement of the device; verbal commands; detected eye movement; biometric input; and / or any combination thereof are optionally used as input corresponding to sub-events defining the event to be recognized.

[0154] Figure 3A portable multifunction device 200 with a touch screen 212 according to some embodiments is shown. The touch screen optionally displays one or more graphics within a user interface (UI) 300. In this embodiment and other embodiments described below, a user can select one or more of these graphics by, for example, making gestures on the graphics using one or more fingers 302 (not drawn to scale in the figure) or one or more styluses 303 (not drawn to scale in the figure). In some embodiments, when the user interrupts contact with one or more graphics, selection of one or more graphics will occur. In some embodiments, gestures optionally include one or more taps, one or more swipes (from left to right, from right to left, up and / or down) and / or rolling of fingers that have been in contact with the device 200 (from right to left, from left to right, up and / or down). In some specific implementations or in some cases, inadvertent contact with a graphic will not select the graphic. For example, when the gesture corresponding to the selection is a tap, a swipe gesture that sweeps over an application icon optionally does not select the corresponding application.

[0155] The device 200 also includes one or more physical buttons, such as a "home" or menu button 304. As previously described, the menu button 304 is used to navigate to any application 236 in a set of applications executing on the device 200. Alternatively, in some embodiments, the menu button is implemented as a soft key in a GUI displayed on the touch screen 212.

[0156] In some embodiments, the device 200 includes a touch screen 212, a menu button 304, a push button 306 for powering the device on / off and for locking the device, one or more volume adjustment buttons 308, a user identity module (SIM) card slot 310, an earphone jack 312, and a docking / charging external port 224. The push button 306 is optionally used to turn the device on / off by pressing the button and keeping the button in a pressed state for a predefined time interval; lock the device by pressing the button and releasing the button before the predefined time interval passes; and / or unlock the device or initiate an unlocking process. In an alternative embodiment, the device 200 also accepts speech input for activating or deactivating certain functions through a microphone 213. The device 200 also optionally includes one or more contact strength sensors 265 for detecting the strength of contact on the touch screen 212, and / or one or more tactile output generators 267 for generating tactile output for a user of the device 200.

[0157] Figure 44 is a block diagram of an exemplary multi-function device with a display and a touch-sensitive surface according to some embodiments. Device 400 does not have to be portable. In some embodiments, device 400 is a laptop computer, a desktop computer, a tablet computer, a multimedia player device, a navigation device, an educational device (such as a children's learning toy), a game system, or a control device (e.g., a home controller or an industrial controller). Device 400 typically includes one or more processing units (CPUs) 410, one or more network or other communication interfaces 460, a memory 470, and one or more communication buses 420 for interconnecting these components. Communication bus 420 optionally includes circuits (sometimes referred to as chipsets) that interconnect system components and control communications between system components. Device 400 includes an input / output (I / O) interface 430 with a display 440, which is typically a touch screen display. I / O interface 430 also optionally includes a keyboard and / or mouse (or other pointing device) 450 and a touchpad 455, a tactile output generator 457 for generating tactile output on device 400 (e.g., similar to the above reference Figure 2A The one or more tactile output generators 267 described above), sensors 459 (e.g., optical sensors, acceleration sensors, proximity sensors, touch sensors, and / or contact intensity sensors (similar to those described above with reference to Figure 2A The memory 470 may include one or more storage devices located away from the CPU 410. In some embodiments, the memory 470 stores information related to the portable multifunction device 200 ( Figure 2A ) or a subset thereof. In addition, memory 470 optionally stores additional programs, modules, and data structures not present in memory 202 of portable multifunction device 200. For example, memory 470 of device 400 optionally stores drawing module 480, presentation module 482, word processing module 484, website creation module 486, disk editing module 488, and / or spreadsheet module 490, while portable multifunction device 200( Figure 2A )'s memory 202 optionally does not store these modules.

[0158] Figure 4Each of the above-mentioned elements in some examples is stored in one or more previously mentioned memory devices. Each module in the above-mentioned modules corresponds to an instruction set for performing the above-mentioned functions. The above-mentioned modules or programs (e.g., instruction sets) do not have to be implemented as independent software programs, processes or modules, so the various subsets of these modules are combined or otherwise rearranged in various embodiments. In some embodiments, memory 470 stores a subset of the above-mentioned modules and data structures. In addition, memory 470 stores additional modules and data structures not described above.

[0159] Attention is now turned to embodiments of a user interface that may be implemented on, for example, portable multifunction device 200.

[0160] Figure 5A An exemplary user interface of an application menu on portable multifunction device 200 is shown according to some embodiments. A similar user interface is implemented on device 400. In some embodiments, user interface 500 includes the following elements, or a subset or superset thereof:

[0161] one or more signal strength indicators 502 of one or more wireless communications such as cellular signals and Wi-Fi signals;

[0162] Time 504;

[0163] Bluetooth indicator 505;

[0164] Battery status indicator 506;

[0165] A tray 508 with icons for commonly used applications, such as:

[0166] o an icon 516 labeled “Phone” of the phone module 238, which optionally includes an indicator 514 of the number of missed calls or voice messages;

[0167] o an icon 518 labeled “Mail” of the email client module 240 , which optionally includes an indicator 510 of the number of unread emails;

[0168] o an icon 520 labeled "Browser" of the browser module 247; and

[0169] o an icon 522 labeled “iPod” of the video and music player module 252 (also referred to as the iPod (trademark of Apple Inc.) module 252); and

[0170] Icons for other apps, such as:

[0171] o Icon 524 labeled "Messages" of IM module 241;

[0172] o Icon 526 labeled “Calendar” of calendar module 248;

[0173] o Icon 528 labeled "Photos" of the image management module 244;

[0174] o Icon 530 labeled “Camera” of camera module 243;

[0175] o Icon 532 labeled “Online Video” of the online video module 255;

[0176] o Icon 534 labeled “Stock Market” of the Stock Market widget 249-2;

[0177] o Icon 536 labeled “Map” of the map module 254;

[0178] o Icon 538 labeled “Weather” of the Weather widget 249-1;

[0179] o Icon 540 labeled “Clock” of the Alarm Clock widget 249-4;

[0180] o An icon 542 labeled “Fitness Support” of the fitness support module 242;

[0181] o Icon 544 labeled "Notepad" of the Notepad module 253; and

[0182] o An icon 546 labeled “Settings” for a settings application or module that provides access to settings for the device 200 and its various applications 236 .

[0183] It should be noted that Figure 5A The icon labels shown in are exemplary only. For example, icon 522 of video and music player module 252 is optionally labeled "Music" or "Music Player". Other labels are optionally used for various application icons. In some embodiments, the label of the corresponding application icon includes the name of the application corresponding to the corresponding application icon. In some embodiments, the label of a particular application icon is different from the name of the application corresponding to the particular application icon.

[0184] Figure 5B A touch-sensitive surface 551 (eg, touch screen display 212) is shown having a touch-sensitive surface 551 (eg, Figure 4 tablet or touch pad 455) of the device (e.g., Figure 4Device 400 also optionally includes one or more contact intensity sensors (e.g., one or more of sensors 457) for detecting intensity of contacts on touch-sensitive surface 551 and / or one or more tactile output generators 459 for generating tactile output for a user of device 400.

[0185] Although some of the examples that follow will be given with reference to input on a touch screen display 212 (where a touch-sensitive surface and a display are combined), in some embodiments, the device detects input on a touch-sensitive surface that is separate from the display, such as Figure 5B In some embodiments, the touch-sensitive surface (e.g., Figure 5B 551) has a main axis (e.g., Figure 5B 553) corresponding to the principal axis (for example, Figure 5B According to these embodiments, the device detects a position corresponding to a corresponding position on the display (e.g., Figure 5B , 560 corresponds to 568 and 562 corresponds to 570) at a contact with touch-sensitive surface 551 (eg, Figure 5B Thus, on a touch-sensitive surface (e.g., Figure 5B 551) and a display of a multi-function device (e.g., Figure 5B When the user input detected by the device on the touch-sensitive surface (e.g., contacts 560 and 562 and their movement) is separated, the device is used to manipulate the user interface on the display. It should be understood that similar methods are optionally used for other user interfaces described herein.

[0186] In addition, although the following examples are primarily given with reference to finger inputs (e.g., finger contacts, single-finger tap gestures, finger swipe gestures), it should be understood that in some embodiments, one or more of these finger inputs are replaced by input from another input device (e.g., mouse-based input or stylus input). For example, a swipe gesture is optionally replaced by a mouse click (e.g., instead of contact), followed by movement of the cursor along the path of the swipe (e.g., instead of movement of the contact). As another example, a tap gesture is optionally replaced by a mouse click when the cursor is over the location of the tap gesture (e.g., instead of detecting contact, followed by ceasing to detect contact). Similarly, when multiple user inputs are detected simultaneously, it should be understood that multiple computer mice are optionally used simultaneously, or a mouse and finger contact are optionally used simultaneously.

[0187] Fig. 6AAn exemplary personal electronic device 600 is shown. Device 600 includes a body 602. In some embodiments, device 600 includes a body 602 relative to devices 200 and 400 (e.g., Figure 2A-Figure 4 ) or some or all of the features described in the foregoing. In some embodiments, device 600 has a touch-sensitive display screen 604, referred to hereinafter as touch screen 604. As an alternative or in addition to touch screen 604, device 600 has a display and a touch-sensitive surface. As with devices 200 and 400, in some embodiments, touch screen 604 (or touch-sensitive surface) has one or more intensity sensors for detecting the intensity of contact (e.g., touch) being applied. One or more intensity sensors of touch screen 604 (or touch-sensitive surface) provide output data representing the intensity of the touch. The user interface of device 600 responds to touch based on the intensity of the touch, which means that touches of different intensities can invoke different user interface operations on device 600.

[0188] Technology for detecting and processing touch intensity may be found, for example, in related applications: International patent application serial number PCT / US2013 / 040061, filed on May 8, 2013, entitled “Device, Method, and Graphical User Interface for Displaying UserInterface Objects Corresponding to an Application,” and International patent application serial number PCT / US2013 / 069483, filed on November 11, 2013, entitled “Device, Method, and Graphical UserInterface for Transitioning Between Touch Input to Display OutputRelationships,” each of which is hereby incorporated by reference in its entirety.

[0189] In some embodiments, the device 600 has one or more input mechanisms 606 and 608. The input mechanisms 606 and 608, if included, are physical. Examples of physical input mechanisms include push buttons and rotatable mechanisms. In some embodiments, the device 600 has one or more attachment mechanisms. Such attachment mechanisms, if included, may allow the device 600 to be attached to, for example, hats, glasses, earrings, necklaces, shirts, jackets, bracelets, watchbands, bracelets, pants, belts, shoes, wallets, backpacks, etc. These attachment mechanisms allow the user to wear the device 600.

[0190] Figure 6BAn exemplary personal electronic device 600 is shown. In some embodiments, the device 600 includes a Figure 2A , Figure 2B and Figure 4 Some or all of the components described. Device 600 has a bus 612 that operatively couples an I / O portion 614 to one or more computer processors 616 and a memory 618. I / O portion 614 is connected to a display 604, which may have a touch-sensitive component 622 and, optionally, also a touch-intensity sensitive component 624. In addition, I / O portion 614 is connected to a communication unit 630 for receiving application and operating system data using Wi-Fi, Bluetooth, near field communication (NFC), cellular and / or other wireless communication technologies. Device 600 includes input mechanisms 606 and / or 608. For example, input mechanism 606 is a rotatable input device or a depressible input device and a rotatable input device. In some examples, input mechanism 608 is a button.

[0191] In some examples, input mechanism 608 is a microphone. Personal electronic device 600 includes, for example, various sensors such as GPS sensor 632, accelerometer 634, orientation sensor 640 (e.g., compass), gyroscope 636, motion sensor 638, and / or combinations thereof, all of which are operably connected to I / O portion 614.

[0192] The memory 618 of the personal electronic device 600 is a non-transitory computer-readable storage medium for storing computer-executable instructions that, when executed by one or more computer processors 616, for example, cause the computer processors to perform the above-described techniques and processes. The computer-executable instructions are also stored and / or transmitted, for example, in any non-transitory computer-readable storage medium for use by or in conjunction with an instruction execution system, device, or apparatus, such as a computer-based system, a system containing a processor, or other system that can obtain instructions from an instruction execution system, device, or apparatus and execute the instructions. The personal electronic device 600 is not limited to Figure 6B components and configurations, but may include other components or additional components in a variety of configurations.

[0193] As used herein, the term “indicative representation” refers to, for example, a display in device 200, 400, and / or 600 ( Figure 2A , Figure 4 and Figure 6A-6B ) is a user-interactive graphical user interface object displayed on a display screen of a computer. For example, an image (e.g., an icon), a button, and text (e.g., a hyperlink) each constitute an affordance.

[0194] As used herein, the term "focus selector" refers to an input element used to indicate the current portion of a user interface with which a user is interacting. In some implementations that include a cursor or other position marker, the cursor acts as a "focus selector" such that when the cursor is over a particular user interface element (e.g., a button, window, slider, or other user interface element), a focus selector is displayed on a touch-sensitive surface (e.g., Figure 4 Touchpad 455 or Figure 5B In the event that an input (e.g., a press input) is detected on the touch-sensitive surface 551 in FIG. 5 , the particular user interface element is adjusted according to the detected input. In the case that a touch screen display (e.g., Figure 2A A touch sensitive display system 212 or Figure 5A In some implementations of the touch screen 212 in FIG. 1 , a contact detected on the touch screen acts as a “focus selector” such that when an input (e.g., a press input by the contact) is detected at the location of a particular user interface element (e.g., a button, window, slider, or other user interface element) on the touch screen display, the particular user interface element is adjusted in accordance with the detected input. In some implementations, the focus moves from one area of ​​the user interface to another area of ​​the user interface without corresponding movement of a cursor or movement of a contact on the touch screen display (e.g., by using a tab key or arrow keys to move the focus from one button to another); in these implementations, the focus selector moves in accordance with the movement of the focus between different areas of the user interface. Regardless of the specific form the focus selector takes, the focus selector is typically a user interface element (or contact on the touch screen display) that is controlled by the user to deliver the user's intended interaction with the user interface (e.g., by indicating to the device the element of the user interface with which the user desires to interact). For example, when a press input is detected on a touch-sensitive surface (e.g., a trackpad or touch screen), the position of a focus selector (e.g., a cursor, contact, or selection box) over a corresponding button will indicate that the user intends to activate the corresponding button (rather than other user interface elements shown on the device display).

[0195] As used in the specification and claims, the term "characteristic intensity" of a contact refers to a characteristic of a contact based on one or more intensities of the contact. In some embodiments, the characteristic intensity is based on multiple intensity samples. The characteristic intensity is optionally based on a predefined number of intensity samples or a set of intensity samples collected during a predetermined time period (e.g., 0.05 seconds, 0.1 seconds, 0.2 seconds, 0.5 seconds, 1 second, 2 seconds, 5 seconds, 10 seconds) relative to a predefined event (e.g., after contact is detected, before contact is detected to be lifted off, before or after contact starts to move, before contact ends, before or after contact intensity is detected to increase, and / or before or after contact intensity is detected to decrease). The characteristic intensity of a contact is optionally based on one or more of the following: the maximum value of the contact intensity, the mean value of the contact intensity, the average value of the contact intensity, the value at the top 10% of the contact intensity, the half-maximum value of the contact intensity, the 90% maximum value of the contact intensity, etc. In some embodiments, the duration of the contact is used when determining the characteristic intensity (e.g., when the characteristic intensity is the average value of the intensity of the contact over time). In some embodiments, the feature strength is compared to a set of one or more strength thresholds to determine whether the user has performed an operation. For example, the set of one or more strength thresholds includes a first strength threshold and a second strength threshold. In this example, a contact whose feature strength does not exceed the first threshold results in a first operation, a contact whose feature strength exceeds the first strength threshold but does not exceed the second strength threshold results in a second operation, and a contact whose feature strength exceeds the second threshold results in a third operation. In some embodiments, a comparison between the feature strength and one or more thresholds is used to determine whether to perform one or more operations (e.g., whether to perform the corresponding operation or give up performing the corresponding operation), rather than to determine whether to perform the first operation or the second operation.

[0196] In some embodiments, a portion of a gesture is identified for determining characteristic strength. For example, a touch-sensitive surface receives a continuous swipe contact that transitions from a starting position and reaches an end position, where the strength of the contact increases. In this example, the characteristic strength of the contact at the end position is based only on a portion of the continuous swipe contact, rather than the entire swipe contact (e.g., the swipe contact is only located at the portion at the end position). In some embodiments, a smoothing algorithm is applied to the strength of the swipe contact before determining the characteristic strength of the contact. For example, a smoothing algorithm optionally includes one or more of the following: an unweighted sliding average smoothing algorithm, a triangular smoothing algorithm, a median filter smoothing algorithm, and / or an exponential smoothing algorithm. In some cases, these smoothing algorithms eliminate narrow peaks or depressions in the strength of the swipe contact to achieve the purpose of determining the characteristic strength.

[0197] The intensity of a contact on a touch-sensitive surface is characterized relative to one or more intensity thresholds such as a contact detection intensity threshold, a light press intensity threshold, a deep press intensity threshold, and / or one or more other intensity thresholds. In some embodiments, the light press intensity threshold corresponds to an intensity at which the device will perform an operation typically associated with clicking a button of a physical mouse or touchpad. In some embodiments, the deep press intensity threshold corresponds to an intensity at which the device will perform an operation different from the operation typically associated with clicking a button of a physical mouse or touchpad. In some embodiments, when a contact having a characteristic intensity below the light press intensity threshold (e.g., and above a nominal contact detection intensity threshold, contacts lower than the nominal contact detection intensity threshold are no longer detected) is detected, the device will move the focus selector according to the movement of the contact on the touch-sensitive surface without performing an operation associated with the light press intensity threshold or the deep press intensity threshold. Generally speaking, unless otherwise stated, these intensity thresholds are consistent between different groups of user interface illustrations.

[0198] An increase in contact feature intensity from an intensity below a light press intensity threshold to an intensity between the light press intensity threshold and the deep press intensity threshold is sometimes referred to as a "light press" input. An increase in contact feature intensity from an intensity below a deep press intensity threshold to an intensity above the deep press intensity threshold is sometimes referred to as a "deep press" input. An increase in contact feature intensity from an intensity below a contact detection intensity threshold to an intensity between the contact detection intensity threshold and the light press intensity threshold is sometimes referred to as detecting contact on the touch surface. A decrease in contact feature intensity from an intensity above the contact detection intensity threshold to an intensity below the contact detection intensity threshold is sometimes referred to as detecting contact lifted from the touch surface. In some embodiments, the contact detection intensity threshold is zero. In some embodiments, the contact detection intensity threshold is greater than zero.

[0199] In some embodiments described herein, one or more operations are performed in response to detecting a gesture that includes a corresponding press input or in response to detecting a corresponding press input performed using a corresponding contact (or multiple contacts), wherein the corresponding press input is detected at least in part based on detecting that the strength of the contact (or multiple contacts) increases to above a press input strength threshold. In some embodiments, the corresponding operation is performed in response to detecting that the strength of the corresponding contact increases to above the press input strength threshold (e.g., a "down stroke" of the corresponding press input). In some embodiments, the press input includes an increase in the strength of the corresponding contact to above the press input strength threshold and the strength of the contact subsequently decreases to below the press input strength threshold, and the corresponding operation is performed in response to detecting that the strength of the corresponding contact subsequently decreases to below the press input threshold (e.g., an "up stroke" of the corresponding press input).

[0200] In some embodiments, the device employs intensity hysteresis to avoid unexpected inputs sometimes referred to as "jitter," where the device defines or selects a hysteresis intensity threshold that has a predefined relationship to a press input intensity threshold (e.g., the hysteresis intensity threshold is X intensity units lower than the press input intensity threshold, or the hysteresis intensity threshold is 75%, 90%, or some reasonable proportion of the press input intensity threshold). Thus, in some embodiments, a press input includes an increase in the intensity of the corresponding contact to above the press input intensity threshold and a subsequent decrease in the intensity of the contact to below the hysteresis intensity threshold corresponding to the press input intensity threshold, and a corresponding operation is performed in response to detecting that the intensity of the corresponding contact subsequently decreases to below the hysteresis intensity threshold (e.g., an "upward stroke" of the corresponding press input). Similarly, in some embodiments, a press input is detected only when the device detects that the contact intensity increases from an intensity equal to or below the hysteresis intensity threshold to an intensity equal to or above the press input intensity threshold and, optionally, that the contact intensity subsequently decreases to an intensity equal to or below the hysteresis intensity, and a corresponding operation is performed in response to detecting the press input (e.g., an increase in contact intensity or a decrease in contact intensity, depending on the circumstances).

[0201] For ease of explanation, optionally, a description of an operation performed in response to a press input associated with a press input intensity threshold or in response to a gesture including a press input is triggered in response to detecting any of the following: contact intensity increasing to above the press input intensity threshold, contact intensity increasing from an intensity below a hysteresis intensity threshold to an intensity above the press input intensity threshold, contact intensity decreasing to below the press input intensity threshold, and / or contact intensity decreasing to below a hysteresis intensity threshold corresponding to the press input intensity threshold. Additionally, in examples where an operation is described as being performed in response to detecting that the intensity of the contact decreases below the press input intensity threshold, the operation is optionally performed in response to detecting that the intensity of the contact decreases below a hysteresis intensity threshold that corresponds to and is less than the press input intensity threshold.

[0202] 3. Digital Assistant System

[0203] Fig. 7A A block diagram of a digital assistant system 700 according to various examples is shown. In some examples, the digital assistant system 700 is implemented on a stand-alone computer system. In some examples, the digital assistant system 700 is distributed across multiple computers. In some examples, some of the modules and functions of the digital assistant are divided into a server portion and a client portion, where the client portion is located on one or more user devices (e.g., device 104, device 122, device 200, device 400, or device 600) and communicates with the server portion (e.g., server system 108) via one or more networks, such as Figure 1 In some examples, the digital assistant system 700 is Figure 1 Specific implementation of the server system 108 (and / or DA server 106) shown in . It should be noted that the digital assistant system 700 is only one example of a digital assistant system, and the digital assistant system 700 has more or fewer components than shown, combines two or more components, or may have a different configuration or layout of components. Fig. 7A The various components shown in the drawings are implemented in hardware, software instructions for execution by one or more processors, firmware (including one or more signal processing integrated circuits and / or application specific integrated circuits), or a combination thereof.

[0204] The digital assistant system 700 includes a memory 702, an input / output (I / O) interface 706, a network communication interface 708, and one or more processors 704. These components can communicate with each other via one or more communication buses or signal lines 710.

[0205] In some examples, memory 702 includes non-transitory computer-readable media, such as high-speed random access memory and / or non-volatile computer-readable storage media (e.g., one or more disk storage devices, flash memory devices, or other non-volatile solid-state memory devices).

[0206] In some examples, the I / O interface 706 couples the input / output devices 716 of the digital assistant system 700, such as a display, keyboard, touch screen, and microphone, to the user interface module 722. The I / O interface 706 and the user interface module 722 receive user input (e.g., voice input, keyboard input, touch input, etc.) and process the input accordingly. In some examples, for example, when the digital assistant is implemented on a stand-alone user device, the digital assistant system 700 includes components respectively corresponding to Figure 2A , Figure 4 , FIG. 6A to FIG. 6B Any of the components and I / O communication interfaces described for device 200, device 400, or device 600 in the digital assistant system 700. In some examples, the digital assistant system 700 represents a server portion of a specific implementation of a digital assistant and can interact with a user through a client-side portion located on a user device (e.g., device 104, device 200, device 400, or device 600).

[0207] In some examples, the network communication interface 708 includes one or more wired communication ports 712 and / or wireless transmission and reception circuits 714. One or more wired communication ports receive and send communication signals via one or more wired interfaces such as Ethernet, Universal Serial Bus (USB), FIREWIRE, etc. The wireless circuit 714 receives RF signals and / or optical signals from a communication network and other communication devices and sends RF signals and / or optical signals to a communication network and other communication devices. Wireless communication uses any of a variety of communication standards, protocols and technologies, such as GSM, EDGE, CDMA, TDMA, Bluetooth, Wi-Fi, VoIP, Wi-MAX, or any other suitable communication protocol. The network communication interface 708 enables the digital assistant system 700 to communicate with other devices through a network, such as the Internet, an intranet and / or a wireless network such as a cellular telephone network, a wireless local area network (LAN) and / or a metropolitan area network (MAN).

[0208] In some examples, the memory 702 or the computer-readable storage medium of the memory 702 stores programs, modules, instructions, and data structures, including all or a subset of the following: operating system 718, communication module 720, user interface module 722, one or more application programs 724, and digital assistant module 726. Specifically, the memory 702 or the computer-readable storage medium of the memory 702 stores instructions for performing the above processes. One or more processors 704 execute these programs, modules, and instructions, and read data from or write data to data structures.

[0209] The operating system 718 (e.g., Darwin, RTXC, LINUX, UNIX, iOS, OS X, WINDOWS, or an embedded operating system such as VxWorks) includes various software components and / or drivers for controlling and managing general system tasks (e.g., memory management, storage device control, power management, etc.), and facilitates communication between various hardware, firmware, and software components.

[0210] The communication module 720 facilitates communication between the digital assistant system 700 and other devices through the network communication interface 708. For example, the communication module 720 and the electronic devices (such as Figure 2A , Figure 4 , FIG. 6A to FIG. 6B The communication module 720 also includes various components for processing data received by the wireless circuit 714 and / or the wired communication port 712.

[0211] The user interface module 722 receives commands and / or input from the user (e.g., from a keyboard, touch screen, pointing device, controller, and / or microphone) via the I / O interface 706 and generates user interface objects on the display. The user interface module 722 also prepares output (e.g., speech, sound, animation, text, icon, vibration, tactile feedback, lighting, etc.) and transmits it to the user via the I / O interface 706 (e.g., through a display, audio channel, speaker, touchpad, etc.).

[0212] Applications 724 include programs and / or modules configured to be executed by the one or more processors 704. For example, if the digital assistant system is implemented on a stand-alone user device, applications 724 include user applications such as games, calendar applications, navigation applications, or mail applications. If the digital assistant system 700 is implemented on a server, applications 724 include, for example, resource management applications, diagnostic applications, or scheduling applications.

[0213] The memory 702 also stores a digital assistant module 726 (or the server portion of the digital assistant). In some examples, the digital assistant module 726 includes the following submodules or subsets or supersets thereof: an input / output processing module 728, a speech-to-text (STT) processing module 730, a natural language processing module 732, a dialog flow processing module 734, a task flow processing module 736, a service processing module 738, and a speech synthesis processing module 740. Each of these modules has access to one or more of the following systems or data and models of the digital assistant module 726 or a subset or superset thereof: a knowledge ontology 760, a vocabulary index 744, user data 748, a task flow model 754, a service model 756, and an ASR system 758.

[0214] In some examples, using the processing modules, data, and models implemented in the digital assistant module 726, the digital assistant may perform at least some of the following: converting voice input into text; recognizing user intent expressed in natural language input received from a user; proactively eliciting and obtaining information needed to fully infer user intent (e.g., by disambiguating words, games, intents, etc.); determining a task flow for satisfying the inferred intent; and executing the task flow to satisfy the inferred intent.

[0215] In some examples, such as Figure 7B As shown in FIG. , the I / O processing module 728 can be Fig. 7A The I / O device 716 interacts with the user or through Fig. 7AThe network communication interface 708 in interacts with a user device (e.g., device 104, device 200, device 400, or device 600) to obtain user input (e.g., voice input) and provide a response to the user input (e.g., as voice output). The I / O processing module 728 optionally obtains context information associated with the user input from the user device along with or shortly after receiving the user input. The context information includes user-specific data, vocabulary, and / or preferences related to the user input. In some examples, the context information also includes the software state and hardware state of the user device when the user request is received, and / or information related to the user's surrounding environment when the user request is received. In some examples, the I / O processing module 728 also sends follow-up questions related to the user request to the user and receives answers from the user. When the user request is received by the I / O processing module 728 and the user request includes voice input, the I / O processing module 728 forwards the voice input to the STT processing module 730 (or speech recognizer) for speech-to-text conversion.

[0216] The STT processing module 730 includes one or more ASR systems 758. The one or more ASR systems 758 can process the speech input received by the I / O processing module 728 to produce a recognition result. Each ASR system 758 includes a front-end speech preprocessor. The front-end speech preprocessor extracts representative features from the speech input. For example, the front-end speech preprocessor performs Fourier transform on the speech input to extract the spectral features that characterize the speech input as a sequence of representative multidimensional vectors. In addition, each ASR system 758 includes one or more speech recognition models (e.g., acoustic models and / or language models) and implements one or more speech recognition engines. Examples of speech recognition models include hidden Markov models, Gaussian mixture models, deep neural network models, n-gram language models, and other statistical models. Examples of speech recognition engines include engines based on dynamic time warping and engines based on weighted finite state transformers (WFST). One or more speech recognition models and one or more speech recognition engines are used to process the extracted representative features of the front-end speech preprocessor to produce intermediate recognition results (e.g., phonemes, phoneme strings and sub-words), and finally produce text recognition results (e.g., words, word strings, or symbol sequences). In some examples, the voice input is at least partially processed by a third-party service or processed on the user's device (e.g., device 104, device 200, device 400 or device 600) to produce recognition results. Once the STT processing module 730 generates a recognition result containing a text string (e.g., a word, or a sequence of words, or a sequence of symbols), the recognition result is transmitted to the natural language processing module 732 for intention inference. In some examples, the STT processing module 730 generates multiple candidate text representations of the voice input. Each candidate text representation is a sequence of words or symbols corresponding to the voice input. In some examples, each candidate text representation is associated with a voice recognition confidence score. Based on the speech recognition confidence score, the STT processing module 730 ranks the candidate text representations and provides the n best (e.g., the n highest ranked) candidate text representations to the natural language processing module 732 for intent inference, where n is a predetermined integer greater than zero. For example, in one example, only the highest ranked (n=1) candidate text representation is delivered to the natural language processing module 732 for intent inference. For another example, the 5 highest ranked (n=5) candidate text representations are delivered to the natural language processing module 732 for intent inference.

[0217] Further details regarding speech-to-text processing are described in U.S. Utility Patent Application Serial No. 13 / 236,942, filed on September 20, 2011, entitled “Consolidating Speech Recognition Results,” the entire disclosure of which is incorporated herein by reference.

[0218] In some examples, the STT processing module 730 includes a vocabulary of recognizable words and / or accesses the vocabulary via the speech-to-alphabet conversion module 731. Each vocabulary word is associated with one or more candidate pronunciations of the word represented in the speech recognition phonetic alphabet. Specifically, the vocabulary of recognizable words includes words associated with multiple candidate pronunciations. For example, the vocabulary includes words associated with and In some examples, the candidate pronunciations for the words are associated with candidate pronunciations for the word "tomato". Additionally, vocabulary words are associated with custom candidate pronunciations based on previous speech input from the user. Such custom candidate pronunciations are stored in the STT processing module 730 and associated with a particular user via a user profile on the device. In some examples, the candidate pronunciations for the words are determined based on the spelling of the words and one or more linguistic and / or phonetic rules. In some examples, the candidate pronunciations are manually generated, for example, based on a known standard pronunciation.

[0219] In some examples, candidate pronunciations are ranked based on their prevalence. Ranked higher than Because the former is a more common pronunciation (e.g., among all users, for users in a particular geographic region, or for any other suitable subset of users). In some examples, the candidate pronunciations are ranked based on whether the candidate pronunciation is a custom candidate pronunciation associated with the user. For example, a custom candidate pronunciation is ranked higher than a standard candidate pronunciation. This can be used to identify proper nouns with unique pronunciations that deviate from the standard pronunciation. In some examples, the candidate pronunciation is associated with one or more speech features such as geographic origin, country, or ethnicity. For example, the candidate pronunciation associated with the United States, while the candidate pronunciation associated with the United Kingdom. In addition, the ranking of the candidate pronunciations is based on one or more characteristics of the user (e.g., geographic origin, country, ethnicity, etc.) in a user profile stored on the device. For example, it may be determined from the user profile that the user is associated with the United States. Based on the user being associated with the United States, the candidate pronunciations are ranked (associated with the United States) Comparable candidate pronunciations (associated with the United Kingdom) is ranked higher. In some examples, one of the ranked candidate pronunciations may be selected as the predicted pronunciation (eg, the most likely pronunciation).

[0220] When speech input is received, the STT processing module 730 is used to determine (e.g., using an acoustic model) a phoneme corresponding to the speech input and then attempts to determine (e.g., using a language model) a word that matches the phoneme. For example, if the STT processing module 730 first identifies a phoneme sequence corresponding to a portion of the speech input It can then determine based on the lexical index 744 that the sequence corresponds to the word "tomato".

[0221] In some examples, the STT processing module 730 uses fuzzy matching techniques to determine words in an utterance. Thus, for example, the STT processing module 730 determines the phoneme sequence corresponds to the word "tomato", even though this particular phoneme sequence is not a candidate phoneme sequence for that word.

[0222] The digital assistant's natural language processing module 732 ("natural language processor") obtains the n best candidate text representations ("word sequences" or "symbol sequences") generated by the STT processing module 730 and attempts to associate each candidate text representation with one or more "executable intents" recognized by the digital assistant. An "executable intent" (or "user intent") represents a task that can be executed by the digital assistant and can have an associated task flow implemented in the task flow model 754. The associated task flow is a series of programmed actions and steps taken by the digital assistant to perform the task. The scope of the digital assistant's capabilities depends on the number and variety of task flows that have been implemented and stored in the task flow model 754, or in other words, on the number and variety of "executable intents" recognized by the digital assistant. However, the effectiveness of the digital assistant also depends on the assistant's ability to infer the correct "one or more executable intents" from user requests expressed in natural language.

[0223] In some examples, in addition to the sequence of words or symbols obtained from the STT processing module 730, the natural language processing module 732 also receives context information associated with the user request, for example, from the I / O processing module 728. The natural language processing module 732 optionally uses the context information to clarify, supplement and / or further qualify the information contained in the candidate text representation received from the STT processing module 730. The context information includes, for example, user preferences, hardware and / or software state of the user's device, sensor information collected before, during, or shortly after the user's request, previous interactions (e.g., conversations) between the digital assistant and the user, and the like. As described herein, in some examples, the context information is dynamic and changes with the time, location, content, and other factors of the conversation.

[0224] In some examples, natural language processing is based on, for example, knowledge ontology 760. Knowledge ontology 760 is a hierarchical structure containing many nodes, each node representing an "executable intent" or an "attribute" related to one or more of the "executable intent" or other "attributes". As described above, an "executable intent" represents a task that the digital assistant can perform, that is, the task is "executable" or can be performed. "Attributes" represent parameters associated with sub-aspects of an executable intent or another attribute. The connection between executable intent nodes and attribute nodes in knowledge ontology 760 defines how the parameters represented by the attribute nodes are subordinate to the tasks represented by the executable intent nodes.

[0225] In some examples, the knowledge ontology 760 is composed of executable intent nodes and attribute nodes. In the knowledge ontology 760, each executable intent node is directly connected to or connected to one or more attribute nodes through one or more intermediate attribute nodes. Similarly, each attribute node is directly connected to or connected to one or more executable intent nodes through one or more intermediate attribute nodes. For example, Figure 7C As shown, ontology 760 includes a "restaurant reservation" node (i.e., an executable intent node). The attribute nodes "restaurant", "date / time" (for reservations), and "party size" are all directly connected to the executable intent node (i.e., the "restaurant reservation" node).

[0226] In addition, the attribute nodes "cuisine", "price range", "telephone number", and "location" are child nodes of the attribute node "restaurant", and are all connected to the "restaurant reservation" node (i.e., executable intent node) through the intermediate attribute node "restaurant". Figure 7C As shown, ontology 760 also includes a "set reminder" node (i.e., another executable intent node). The attribute nodes "date / time" (for setting reminders) and "subject" (for reminders) are both connected to the "set reminder" node. Since the attribute "date / time" is related to both the task of making a restaurant reservation and the task of setting a reminder, the attribute node "date / time" is connected to both the "restaurant reservation" node and the "set reminder" node in ontology 760.

[0227] Executable intent nodes, together with their linked attribute nodes, are described as "domains". In this discussion, each domain is associated with a corresponding executable intent and refers to a set of nodes (and the relationships between these nodes) associated with a particular executable intent. For example, Figure 7CThe ontology 760 shown in includes an example of a restaurant reservation domain 762 and an example of a reminder domain 764 within the ontology 760. The restaurant reservation domain includes an executable intent node "restaurant reservation", attribute nodes "restaurant", "date / time" and "party size", and sub-attribute nodes "cuisine", "price range", "phone number" and "location". The reminder domain 764 includes an executable intent node "set reminder" and attribute nodes "subject" and "date / time". In some examples, the ontology 760 consists of multiple domains. Each domain shares one or more attribute nodes with one or more other domains. For example, in addition to the restaurant reservation domain 762 and the reminder domain 764, the "date / time" attribute node is also associated with many different domains (e.g., itinerary scheduling domain, travel booking domain, movie ticket domain, etc.).

[0228] although Figure 7C Two exemplary domains within ontology 760 are shown, but other domains include, for example, "find a movie," "make a phone call," "find directions," "schedule a meeting," "send a message," and "provide answers to questions," "read a list," "provide navigation instructions," "provide instructions for a task," etc. The "send message" domain is associated with the "send message" executable intent node, and further includes attribute nodes such as "one or more recipients," "message type," and "message body." The attribute node "recipients" is further defined, for example, by child attribute nodes such as "recipient name" and "message address."

[0229] In some examples, ontology 760 includes all domains (and thus executable intents) that a digital assistant can understand and act upon. In some examples, ontology 760 is modified, such as by adding or removing entire domains or nodes, or by modifying the relationships between nodes within ontology 760.

[0230] In some examples, nodes associated with multiple related executable intents are clustered under a "superdomain" in the knowledge ontology 760. For example, the "travel" superdomain includes a cluster of attribute nodes and executable intent nodes related to travel. Executable intent nodes related to travel include "ticket booking", "hotel booking", "car rental", "get directions", "find points of interest", and so on. Executable intent nodes under the same superdomain (e.g., the "travel" superdomain) have multiple shared attribute nodes. For example, the executable intent nodes for "ticket booking", "hotel booking", "car rental", "get directions", and "find points of interest" share one or more of the attribute nodes "starting location", "destination", "departure date / time", "arrival date / time", and "party size".

[0231] In some examples, each node in the knowledge ontology 760 is associated with a set of words and / or phrases related to the attribute or executable intent represented by the node. The corresponding set of words and / or phrases associated with each node is a so-called "vocabulary" associated with the node. The corresponding set of words and / or phrases associated with each node is stored in the vocabulary index 744 associated with the attribute or executable intent represented by the node. For example, return Figure 7B , the vocabulary associated with the node of the "restaurant" attribute includes words such as "food", "drinks", "cuisine", "hunger", "eat", "pizza", "fast food", "meals", etc. For another example, the vocabulary associated with the node of the "make phone call" executable intent includes words and phrases such as "call", "make a phone call", "dial", "talk to...", "call the number", "call", etc. The vocabulary index 744 optionally includes words and phrases in different languages.

[0232] The natural language processing module 732 receives candidate text representations (e.g., one or more text strings or one or more symbol sequences) from the STT processing module 730, and for each candidate representation, determines which nodes the words in the candidate text representation involve. In some examples, if a word or phrase in the candidate text representation is found to be associated with one or more nodes in the knowledge ontology 760 (via the vocabulary index 744), the word or phrase "triggers" or "activates" those nodes. Based on the number and / or relative importance of the activated nodes, the natural language processing module 732 selects one of the executable intents as the task that the user intends the digital assistant to perform. In some examples, the domain with the most "triggered" nodes is selected. In some examples, the domain with the highest confidence (e.g., based on the relative importance of its various triggered nodes) is selected. In some examples, the domain is selected based on a combination of the number and importance of the triggered nodes. In some examples, additional factors are also considered in the process of selecting nodes, such as whether the digital assistant has previously correctly interpreted similar requests from the user.

[0233] User data 748 includes user-specific information, such as user-specific vocabulary, user preferences, user address, user's default second language, user's contact list, and other short-term or long-term information for each user. In some examples, natural language processing module 732 uses user-specific information to supplement the information contained in the user input to further qualify the user's intent. For example, for a user request to "invite my friends to my birthday party," natural language processing module 732 can access user data 748 to determine who "friends" are and when and where the "birthday party" will be held, without requiring the user to explicitly provide such information in their request.

[0234] It should be recognized that in some examples, the natural language processing module 732 is implemented using one or more machine learning mechanisms (e.g., neural networks). Specifically, the one or more machine learning mechanisms are configured to receive candidate text representations and context information associated with the candidate text representations. Based on the candidate text representations and the associated context information, the one or more machine learning mechanisms are configured to determine an intent confidence score based on a set of candidate executable intents. The natural language processing module 732 may select one or more candidate executable intents from a set of candidate executable intents based on the determined intent confidence scores. In some examples, a knowledge ontology (e.g., knowledge ontology 760) is also utilized to select one or more candidate executable intents from a set of candidate executable intents.

[0235] Additional details of searching knowledge ontologies based on symbolic strings are described in U.S. Utility Patent Application Serial No. 12 / 341,743, filed on December 22, 2008, entitled “Method and Apparatus for Searching Using An Active Ontology,” the entire disclosure of which is incorporated herein by reference.

[0236] In some examples, once the natural language processing module 732 identifies an executable intent (or domain) based on a user request, the natural language processing module 732 generates a structured query to represent the identified executable intent. In some examples, the structured query includes parameters for one or more nodes within the domain of the executable intent, and at least some of the parameters are filled with specific information and requirements specified in the user request. For example, the user says "help me reserve a seat at 7pm at the sushi restaurant." In this case, the natural language processing module 732 is able to correctly identify the executable intent as "restaurant reservation" based on the user input. According to the knowledge ontology, the structured query of the "restaurant reservation" domain includes parameters such as {cuisine}, {time}, {date}, {party number}, etc. In some examples, based on voice input and text derived from the voice input using the STT processing module 730, the natural language processing module 732 generates a partial structured query for the restaurant reservation domain, wherein the partial structured query includes parameters {cuisine = "sushi type"} and {time = "7pm"}. However, in this example, the user utterance contains insufficient information to complete the structured query associated with the domain. Therefore, based on the currently available information, other necessary parameters such as {number of party members} and {date} are not specified in the structured query. In some examples, the natural language processing module 732 populates some parameters of the structured query with the received contextual information. For example, in some examples, if the user requests a sushi restaurant "nearby", the natural language processing module 732 populates the {location} parameter in the structured query with the GPS coordinates from the user's device.

[0237] In some examples, the natural language processing module 732 identifies multiple candidate executable intentions for each candidate text representation received from the STT processing module 730. In addition, in some examples, a corresponding structured query is generated (partially or completely) for each identified candidate executable intention. The natural language processing module 732 determines the intent confidence score for each candidate executable intention and ranks the candidate executable intention based on the intent confidence score. In some examples, the natural language processing module 732 transmits the generated one or more structured queries (including any completed parameters) to the task flow processing module 736 ("task flow processor"). In some examples, one or more structured queries for the m best (e.g., the m highest ranked) candidate executable intentions are provided to the task flow processing module 736, where m is a predetermined integer greater than zero. In some examples, one or more structured queries for the m best candidate executable intentions are provided to the task flow processing module 736 together with the corresponding one or more candidate text representations.

[0238] Additional details for inferring user intent based on multiple candidate executable intents determined based on multiple candidate textual representations of speech input are described in U.S. utility model patent application serial number 14 / 298,725, entitled “System and Method for Inferring User Intent From Speech Inputs,” filed on June 6, 2014, the entire disclosure of which is incorporated herein by reference.

[0239] The task flow processing module 736 is configured to receive one or more structured queries from the natural language processing module 732, complete the structured queries (if necessary), and perform the actions required to "complete" the user's final request. In some examples, the various processes necessary to complete these tasks are provided in the task flow model 754. In some examples, the task flow model 754 includes a process for obtaining additional information from the user, and a task flow for performing actions associated with an executable intent.

[0240] As described above, in order to complete the structured query, the task flow processing module 736 needs to initiate an additional dialogue with the user in order to obtain additional information and / or clarify the utterances that may be ambiguous. When it is necessary to conduct such an interaction, the task flow processing module 736 calls the dialogue flow processing module 734 to participate in the dialogue with the user. In some examples, the dialogue flow processor module 734 determines how (and / or when) to request additional information from the user, and receives and processes the user response. The question is provided to the user and the answer is received from the user through the I / O processing module 728. In some examples, the dialogue flow processing module 734 presents the dialogue output to the user via an audible output and / or a visual output, and receives input from the user via a verbal or physical (e.g., click) response. Continuing the above example, when the task flow processing module 736 calls the dialogue flow processing module 734 to determine the "party number" and "date" information for the structured query associated with the domain "restaurant reservation", the dialogue flow processing module 734 generates questions such as "how many people in a row?" and "which day to book?" and passes them to the user. Once the answer is received from the user, the dialog flow processing module 734 fills in the structured query with the missing information, or passes the information to the task flow processing module 736 to complete the missing information based on the structured query.

[0241] Once the task flow processing module 736 has completed the structured query for the executable intent, the task flow processing module 736 begins to execute the final task associated with the executable intent. Therefore, the task flow processing module 736 executes the steps and instructions in the task flow model according to the specific parameters contained in the structured query. For example, the task flow model for the executable intent "restaurant reservation" includes steps and instructions for contacting a restaurant and actually requesting a reservation for a specific party size at a specific time. For example, using a structured query such as: {restaurant reservation, restaurant = ABC Cafe, date = 3 / 12 / 2012, time = 7pm, party size = 5}, the task flow processing module 736 can perform the following steps: (1) Log in to the server of ABC Cafe or such 's restaurant reservation system, (2) entering date, time, and party size information into a form on the website, (3) submitting the form, and (4) creating a calendar entry for the reservation in the user's calendar.

[0242] In some examples, the task flow processing module 736 completes the task requested in the user input or provides the informational answer requested in the user input with the assistance of the service processing module 738 ("service processing module"). For example, the service processing module 738 initiates a phone call, sets a calendar entry, invokes a map search, invokes or interacts with other user applications installed on the user device, and invokes or interacts with third-party services (e.g., restaurant reservation portals, social networking sites, bank portals, etc.) on behalf of the task flow processing module 736. In some examples, the protocols and application programming interfaces (APIs) required for each service are specified by the corresponding service models in the service model 756. The service processing module 738 accesses the appropriate service model for the service and generates a request for the service according to the protocol and API required by the service based on the service model.

[0243] For example, if the restaurant has enabled an online reservation service, the restaurant submits a service model that specifies the necessary parameters for making a reservation and transmits the values ​​of the necessary parameters to the API of the online reservation service. When requested by the task flow processing module 736, the service processing module 738 can use the web address stored in the service model to establish a network connection with the online reservation service and send the necessary parameters for the reservation (e.g., time, date, party size) to the online reservation interface in a format according to the API of the online reservation service.

[0244] In some examples, the natural language processing module 732, the dialog flow processing module 734, and the task flow processing module 736 are used together and repeatedly to infer and define the user's intention, obtain information to further clarify and refine the user's intention, and finally generate a response (i.e., output to the user, or complete the task) to meet the user's intention. The generated response is a dialogue response to the voice input that at least partially meets the user's intention. In addition, in some examples, the generated response is output as a voice output. In these examples, the generated response is sent to a speech synthesis processing module 740 (e.g., a speech synthesizer), where the generated response can be processed to synthesize the dialogue response in the form of speech. In other examples, the generated response is data content related to satisfying the user request in the voice input.

[0245] In an example where the task flow processing module 736 receives multiple structured queries from the natural language processing module 732, the task flow processing module 736 first processes the first structured query of the received structured query in an attempt to complete the first structured query and / or perform one or more tasks or actions represented by the first structured query. In some examples, the first structured query corresponds to the highest-ranked executable intent. In other examples, the first structured query is selected from the structured query received based on the combination of the corresponding speech recognition confidence score and the corresponding intent confidence score. In some examples, if the task flow processing module 736 encounters an error during the processing of the first structured query (e.g., due to the inability to determine the necessary parameters), the task flow processing module 736 may continue to select and process the second structured query corresponding to the executable intent with a lower ranking in the received structured query. For example, the second structured query is selected based on the speech recognition confidence score represented by the corresponding candidate text, the intent confidence score of the corresponding candidate executable intent, the missing necessary parameters in the first structured query, or any combination thereof.

[0246] The speech synthesis processing module 740 is configured to synthesize speech output for presentation to the user. The speech synthesis processing module 740 synthesizes speech output based on the text provided by the digital assistant. For example, the generated dialogue response is in the form of a text string. The speech synthesis processing module 740 converts the text string into an audible speech output. The speech synthesis processing module 740 uses any appropriate speech synthesis technology to generate speech output from the text, including but not limited to: concatenation synthesis, unit selection synthesis, diphone synthesis, domain-specific synthesis, formant synthesis, pronunciation synthesis, synthesis based on hidden Markov model (HMM) and sine wave synthesis. In some examples, the speech synthesis processing module 740 is configured to synthesize individual words based on phoneme strings corresponding to these words. For example, the phoneme string is associated with the words in the generated dialogue response. The phoneme string is stored in the metadata associated with the word. The speech synthesis processing module 740 is configured to directly process the phoneme string in the metadata to synthesize the words in speech form.

[0247] In some examples, instead of (or in addition to) using the speech synthesis processing module 740, speech synthesis is performed on a remote device (e.g., server system 108), and the synthesized speech is sent to the user device for output to the user. For example, this may occur in some implementations where the output of the digital assistant is generated at the server system. And because the server system typically has more processing power or more resources than the user device, it is possible to obtain higher quality speech output than client-side synthesis would achieve.

[0248] Additional details about the digital assistant can be found in U.S. Utility Patent Application No. 12 / 987,982, filed on January 10, 2011, and entitled “Intelligent Automated Assistant,” and U.S. Utility Patent Application No. 13 / 251,088, filed on September 30, 2011, and entitled “Generating and Processing Task Items That Represent Tasks to Perform,” the entire disclosures of which are incorporated herein by reference.

[0249] 4. The process of providing suggested user actions in response to detected anchors

[0250] Embodiments relate to providing suggested user actions via electronic devices. Suggested actions may be provided in response to detecting the occurrence of an anchor. An anchor may be a predefined event that occurs in a user's day. For example, upon waking up in the morning (e.g., user waking up may be an example of an anchor), an electronic device may provide a notification including a suggestion for playing the latest episode of a particular podcast (i.e., a suggested user action), and the user often listens to a particular podcast when waking up in the morning. The occurrence of an anchor may be detected via one or more signals generated by the electronic device (e.g., the alarm function of the device is being executed, the device is converted from "do not disturb mode" to "active mode", etc.). As used herein, an anchor may be a marker or signal of an event in a user's day. The occurrence of an anchor may be detected via monitoring and analyzing electronic signals generated by an electronic device. Based on the user's previous interaction with the device, the occurrence of an anchor may indicate user behavior and / or actions taken in response to an event in a user's day. As a non-limiting example, an anchor may include a user input location of interest (LOI), such as, but not limited to, a user's home, office, gym, airport, shopping mall, etc. Because an anchor may be an event that indicates user behavior, an anchor may be referred to as an anchor event or simply an event throughout the text. The occurrence of an anchor may be referred to as an anchor occurrence.

[0251] Other non-limiting examples of the occurrence of an anchor include a user starting to use an electronic device after an extended period of device idle time, a user ending a workout, a user waking up, a user going to bed, a user pairing a Bluetooth-enabled device with an electronic device, a user starting or completing a calendar event, and the like. Other examples of anchors include a user launching a specific application of a device and / or adopting a specific function of a device. Any event that can be detected via a signature encoded in a signal generated by an electronic device can be an anchor. The signature indicating the occurrence of a specific anchor (e.g., a pattern within an electronic signal generated by a device) can be predetermined, pre-calculated, and / or pre-learned (e.g., via supervised or unsupervised machine learning (ML) methods). Therefore, an electronic device can monitor its various signals and detect the occurrence of a specific anchor by identifying and / or detecting one or more signatures within a signal indicating a specific anchor. For example, the occurrence of a wakeup anchor can be detected via a signal of an electronic device indicating that a user has terminated or otherwise ended the "idle" state of the electronic device (or the alarm of the device rings). Upon detecting the occurrence of an anchor, various embodiments can suggest (via previously generated training data) one or more user actions associated with the occurrence of the anchor.

[0252] A user action (i.e., an action) may include invoking, executing, or otherwise initiating a specific application, capability, function, or command enabled via an electronic device. As non-limiting examples, a user action may include, but is not limited to, playing specific audio / video content (e.g., a podcast, a music playlist, an audiobook, a concert, a television series, a movie, etc.), launching an application installed on a device (e.g., a fitness application, a meditation application, a ride-sharing application, a food delivery application, a social networking application, etc.), sending an electronic message (e.g., an email, SMS, Twitter, etc.) to another user, a user group, or a social network, turning on / off a function of a device (e.g., turning off / on airplane mode of a device), updating various settings and / or configurations of a device, creating a calendar event, etc.

[0253] When providing a suggested action, various embodiments may provide a notification (e.g., a "pop-up" notification) indicating the suggested action. The notification may be interactive, in that the user may initiate an action via an interactive selection of the notification. Upon selecting the notification, the suggested action may be performed by the device and / or the execution of the action may be initiated. Such suggested action notifications include, but are not limited to, pop-up notifications, toast notifications, passive pop-up notifications, fast food notifications, bubble notifications, or any other such notifications. In at least some embodiments, the suggested action may be automatically executed without the user interacting with the notification. For example, the user may adopt a setting that enables the suggested action to be automatically executed. As described throughout, the suggested action for any given anchor may not be unique. For example, multiple actions may be suggested to the user when the user wakes up. The context of the anchor may affect the determination of which actions are suggested. For example, the occurrence of waking up the anchor during the weekday may result in a suggestion to play a podcast, while the occurrence of waking up the anchor on the weekend may result in a suggestion to start fitness via a fitness application installed on the device.

[0254] In various embodiments, one or more machine learning (ML) models may be trained to learn various statistical correlations (or associations) between the occurrence of a particular anchor and an action initiated by a user (e.g., an innovation in a device function or capability). Any of the various ML models employed by various embodiments may be collectively referred to herein as an "anchor model" because these models are trained to detect the occurrence of various anchors and suggest one or more actions that are statistically likely to be initiated by the user in response to the occurrence (or imminent occurrence) of the anchor. Therefore, various embodiments include the training of anchor models, and the use of trained anchor models to enhance the user's experience (UX) of adopting the device. That is, once deployed to the device, the trained anchor model significantly enhances the performance of the device because the anchor model is used to predict and suggest one or more actions that the user may take in response to various anchors (or events) that occur throughout the user's day. Supervised or unsupervised ML may be used to train the model to determine the context (e.g., contextual features) in which the anchor occurs, and to establish, customize, and / or selectively target the suggested action based on the context of the anchor.

[0255] More specifically, embodiments may detect, determine and / or otherwise identify various contextual features of anchor occurrences. Embodiments may employ detected contextual features of anchor occurrences and customize the one or more suggested actions to the context of the occurrence of the anchor. That is, in addition to the occurrence of the anchor, one or more signatures within the monitored signal may provide contextual information (e.g., contextual features) associated with a particular occurrence of the anchor. The one or more suggested actions provided may be based on contextual information. The contextual features of the occurrence of the anchor (e.g., user wake-up) may include a calendar date, a day of the week, the user's location, the time of the day of the anchor occurrence, and any other such contextual information. Various suggested actions may be provided and customized based on such contextual information. For example, if a day of the week is a weekday, or a workday for the user, the one or more suggested actions may include a suggestion to launch a navigation application that provides real-time road traffic information and / or provides information related to commuting to work. In contrast, if a day of the week is a weekend or non-workday for the user, the one or more suggested actions may include different and / or separate actions, such as, but not limited to, suggestions to launch a fitness application or play a podcast. If the time of day when the user wakes up is morning, different suggested actions may be provided, in contrast to the suggested actions if the user is waking up from a nap. In such embodiments, the anchor model may be trained to identify various patterns and / or correlations between anchor occurrences, contextual features of the anchor occurrences, and user responses to the anchor occurrences and contextual features. For anchor types associated with multiple actions, ML may be used to train the model to distinguish the multiple actions based on the context of the anchor occurrences. In some embodiments, even if more than one action type is positively correlated with a particular anchor type, when an occurrence of an anchor type is detected, only the most likely action of the multiple actions is suggested based on the context of the anchor occurrence. That is, even if multiple actions are associated with an anchor, the anchor model may be trained to determine the statistical correlation between the user's possible actions and the anchor occurrence based on the specific contextual features of a particular anchor occurrence.

[0256] In some embodiments, the contextual information of the anchor occurrence may include contextual differences, changes or differences that occur after (or before) the anchor occurs. Such contextual information may indicate the time period (e.g., amount of time) that has passed since the anchor occurs, or the time period until the expected anchor occurs. In such embodiments, the moment when the electronic device provides the one or more suggested actions may be based on contextual information, such as but not limited to the amount of time that has passed since the anchor occurs, or the amount of time until the expected anchor occurs. Therefore, the context of the anchor occurrence can be a context-based context. In some embodiments, the context of the anchor occurrence can be a location-based context. For example, when the user enters a location of interest (LOI), the context can be location-based, for example, an LOI for the user's home or the user's office. In such embodiments using a time-based context, an anchor model can be used to determine the time offset measured from the occurrence time of the anchor (or the expected time of the expected anchor occurrence). The one or more suggested actions can be provided once based on the time offset. For example, the time offset for providing the suggested action can be 15 minutes after the occurrence of the anchor associated with the user's wake-up. Suggested actions (e.g., playing a podcast) can be provided at a time offset (e.g., 15 minutes) after the anchor occurs. That is, a suggestion to play a podcast may be provided at a time offset, as measured after the anchor is detected to occur (e.g., 15 minutes after the user wakes up).

[0257] As described above, one or more anchor models may be employed to detect anchor occurrences and provide suggested actions. In various embodiments, the anchor model may be trained to learn or otherwise predict possible time offsets for providing suggested actions and other contextual features of the anchor. The time offset may be positive (e.g., providing a suggested action after the occurrence of the anchor and after the announcement of the time offset) or negative (e.g., providing a suggested action before the expected occurrence of the anchor). The above example of providing a suggestion to play a podcast 15 minutes after "waking up" the anchor is an example of a positive time offset. An example of a negative time offset includes providing a suggested action that includes calling a ride via a ride-sharing application three hours before the user's scheduled flight departs.

[0258] The detection of an anchor occurrence may trigger the provision of multiple suggested actions. The multiple suggested actions may be provided at separate time offsets. As a non-limiting example, a wakeup anchor may be detected at 7:00 a.m. The anchor may be detected via a signal generated when the user manually terminates (or the electronic device automatically terminates based on the timer setting of the "do not disturb mode") an extended period of "idle time" of the electronic device. At a first positive time offset of 5 minutes (e.g., at 7:05 a.m.), the electronic device provides the user with a first suggested action of playing the latest episode of one of the user's favorite podcasts via a podcast application installed on the device. At a second positive time offset of 60 minutes (e.g., at 8:00 AM), the electronic device provides the user with a second suggested action of using a weather application to check the local weather report. In some embodiments, whether to provide (or not provide) the second suggested action at the second time offset may be based on whether the user initiates the first suggested action. For example, the second suggested action may be provided only if and only if (iff) the user initiates the first suggested action. In other embodiments, the second suggested action may be provided at a second time offset, regardless of whether the user initiates the first suggested action. In other embodiments, the value of the second time offset may be based on whether the user initiates the first suggested action. For example, if the user has not listened to the suggested episode of the podcast, the value of the second time offset can be set to 60 minutes. If the user does listen to the podcast, the value of the second time offset can be set so that the second suggested action is provided when the episode of the podcast is completed or the user otherwise stops listening to the podcast.

[0259] Continuing with this example, regarding the user's day, the user may begin their commute to work by arriving at their bus stop. The bus stop may be a predetermined or "learned" (via ML) LOI, and arriving at the bus stop may be a second anchor in the user's day. Anchor occurrences may be detected via monitoring of GPS signals. Thus, the LOI anchor may be detected when the user arrives at their bus stop. The bus stop may serve as a location-based context for the anchor occurrence. Upon arriving at the LOI (e.g., bus stop), and based on the location context, the user may be provided with a suggested action of checking when the user's bus will arrive via a bus schedule application. As the user approaches their work office via bus, another suggested action may be automatically provided, such as, but not limited to, sending a message (e.g., SMS, email, etc.) to a colleague. Providing the suggested action may depend on the location-based context.

[0260] As another example implementation of a suggested action based on the detection of an anchor, a user may have a flight schedule for later that day. The flight may serve as an anchor or an expected anchor occurrence. Based on the expectation of the flight occurring, one or more suggested actions may be provided prior to the flight via a negative time offset. For example, the flight may be scheduled to take off at 7:00 p.m. Based on the -1:30 time offset (which is "learned" during the training of the anchor model), a suggested action may be provided to the user at 5:30 p.m. The suggested action may include requesting a rideshare service to the airport via a rideshare application installed on the user's device. When the user is near the airport, a location-based contextual suggested action may be provided to the user to query their flight (e.g., using another application). Another time-based contextual suggested action may be provided at 6:55 p.m., with another negative time offset (e.g., providing a message to a family member that the flight is about to take off).

[0261] During the training cycle, the anchor model can be trained via training data generated from the user's interaction with the electronic device. At least a portion of the anchor mark signature within the device signal can be predetermined and / or pre-calculated. At least a portion of the signature can indicate contextual features of the anchor occurrence. Such context signals can also be predetermined and / or pre-calculated. In other embodiments, the anchor annotation signal and / or the context feature annotation signal can be learned and / or determined via training the model. For example, various deep learning (DL) methods can be used to determine the signature. The signature within the electronic signal indicating the occurrence of the anchor can be referred to as the anchor signature, and the signature indicating the contextual features of the anchor occurrence can be referred to as the context signature.

[0262] The training of the anchor model can be performed to be specific to an electronic device and / or a specific user. For example, the training of the anchor model can be performed using training data associated with a specific user and / or a specific electronic device, wherein the training data is not associated with other users and / or other devices. In some embodiments, the computing resources (e.g., processing power, storage devices, and / or memory) of the electronic device can be used to train the model. The training data (e.g., electrical signals) can be generated by a specific device and a device interaction associated with a specific user. There is no need to provide the training data and / or the trained anchor model to other devices. Therefore, the training of the model does not require the use of computing resources separate from the computing resources of the specific electronic device of the specific user. In addition, the privacy of the user can be protected because the training data and / or the trained model does not need to be provided to another electronic device other than the user's device.

[0263] In other embodiments, at least a portion of the training of the model may be offloaded to computing resources associated with other devices. In such embodiments, the user's privacy may still be protected by encrypting the training data and / or the trained model. Prior to offloading, the training data generated by the user's electronic device may be fully encrypted. For example, an encrypted version of the trained model and / or training data may be stored via various user-accessible cloud computing resources. The encryption key may be stored on the user's device so that the trained model and / or training data may be inaccessible to other users and / or other devices. Therefore, even if the training data and / or the trained model may be "backed up" via a cloud service, the user's privacy is still protected.

[0264] In various embodiments, each type of anchor may be trained separately (e.g., wake up, enter LOI, end of last calendar event, end of workout, etc.). In some embodiments, the anchor model may be trained for multiple anchor types in parallel. In other embodiments, multiple anchor types may be trained in series. A combination of parallel training cycles and serial training cycles may be used for multiple anchor types. For each anchor type, the training cycle of the anchor model may include a training data acquisition phase and three learning phases. The training data acquisition phase includes acquiring a sufficient amount of training data for the anchor type. As described below, the first learning phase of training may be referred to as a filtering phase, the second learning phase may be referred to as a selection and / or sorting phase, and the third learning phase may be referred to as a time offset phase.

[0265] Acquiring training data includes acquiring and / or monitoring training signals generated by an electronic device. The training signal includes a statistically significant number of anchor occurrences and a statistically significant number of training actions (e.g., actions performed by the device during the data acquisition phase). The signal may also include a signature indicating contextual features, contextual conditions, or other contextual information associated with the anchor occurrence. The contextual information of each anchor occurrence may be encoded in metadata and associated with and / or included in the data encoding the anchor occurrence. The action may include user-initiated actions that are performed by the device via the user interacting with the device during the training data acquisition phase. Each action in the training data may be classified into one or more action types in a set of enabled action types that can be executed by the device. Therefore, the enabled action type set may be a possible action type set that is enabled and / or executable by the electronic device. For example, the enabled action type set may include playing a podcast set via a podcast application, starting a fitness provided by a fitness application, sending a message via a messenger application, requesting a carpooling service via a carpooling application, and the like.

[0266] After a sufficient amount of training data associated with the anchor type has been acquired by the electronic device during the data acquisition phase, the training of the anchor model of the anchor type may begin in the filtering phase. During the filtering phase, a candidate action type set may be determined and / or generated. The candidate action type set may be determined by filtering the enabled action type set. Because the filtering phase includes filtering the enabled action type set, the candidate action type set may be a subset of the enabled action type set. Filtering the enabled action set may include determining or generating statistical correlations, correspondences, and / or associations between training events and anchor occurrences of training data. Filtering may be performed via statistical correlations. For example, correlations may be used to determine which action types are most highly associated (or related) with the anchor type being trained. As described throughout, various methods may be used to generate correlations and filtering criteria, such as, but not limited to, determining class conditional probabilities or posterior probabilities conditioned on anchor types. In some embodiments, filtering criteria may include entropy-based filtering criteria, local maximum or minimum-based filtering criteria, and the like.

[0267] During the sorting phase, the candidate action type set can be sorted, and at least a portion of the candidate action type can be selected based on the sorting. The sorting and selection of the candidate action type can be based on the contextual conditions or contextual features associated with the anchor occurrence. That is, in response to (or expected) the occurrence of the anchor, the action type that the user is most likely to initiate or perform is highly sorted and selected with a specific contextual condition or contextual feature as a condition. In some embodiments, a classification decision tree is constructed for each action type of the candidate action type set. The node of the classification decision tree for the action type can indicate one or more possible contextual conditions or contextual features associated with the action type or with the action type as a condition and including the corresponding contextual conditions or features of the anchor occurrence. Additionally and / or alternatively, the node can be conditioned by the context of the action type of the tree. For example, the node can be conditioned by the number of times or frequency that the user initiates the action type within a given amount of time (e.g., the previous hour, the previous 12 hours, the previous day, the previous week, etc.). In order to determine the probability or possibility that the user initiates a specific action type, with the anchor occurrence type and one or more contextual conditions as a condition, the nodes of the corresponding classification decision tree can be traversed, and the probability can be calculated based on the whole text. A context-dependent classification decision tree can be used to map multiple action types to a single anchor type based on the context in which the anchor occurs. The multiple action types can be distinguished for a single anchor type via the different contexts in which the anchor occurs. For example, each of a first action type (e.g., play a podcast episode) and a second action type (e.g., start a workout) can be mapped to a single anchor type of wakeup. Thus, the ordering and selection of action types can be based on a classification decision tree.

[0268] During the time offset phase of training for anchor types, time offsets may be determined for one or more action types that may be suggested in response to (or in anticipation of) the occurrence of anchor types. In order to determine the time offset, one or more time distributions may be generated from the training data. The time distribution may include a histogram of the time difference between the occurrence of the anchor and the action initiated by the user. In some embodiments, a separate time distribution may be generated for the action type. The time offset may be determined via one or more statistical features of the distribution (e.g., the mean value of the distribution, the median of the distribution, the mode of the distribution, the variance of the distribution, etc.). For example, when training a model for an anchor type for wakeup, the training data may be used to generate a time histogram of the action type for playing a podcast. The x-axis of the histogram may include intervals of time values ​​between the occurrence of user wakeup and the user initiating the playback of one or more podcast episodes. The y-axis of the histogram may indicate the number of times (or frequencies) that the user initiated the playback of a podcast at the time value and after the occurrence of the user wakeup. In at least some embodiments, a separate time distribution may be generated for each action type and each contextual condition or a combination of contextual conditions. For example, separate histograms (similar to the histograms described above) may be generated for wakeup anchors that occur in the mornings of weekdays (or the user's workdays) and for wakeup anchors that occur in the mornings of weekend days.

[0269] Once the anchor model is fully trained for one or more anchor types, the trained model can be deployed by the electronic device. Once the trained model is deployed, the signal generated by the device can be monitored and / or analyzed. The occurrence of the anchor type trained by the model can be detected and / or identified. As described above, in addition to indicating the anchor occurrence, the signal can indicate one or more contextual conditions of the anchor occurrence. The contextual conditions can be encoded in the metadata associated with the anchor occurrence. In response to and / or based on the detected anchor occurrence, the model can be used to determine and provide one or more suggested actions and the time offset of each of the one or more suggested actions. The one or more suggested actions and the associated time offset can be based on the contextual conditions of the anchor occurrence. At the time offset of the suggested action measured from the anchor occurrence, the electronic device can provide a notification indicating the suggested action to the user. The user can be enabled to select the notification, and when the user selects, the electronic device can perform the selected suggested action. As discussed throughout, the model can be used to detect anchor occurrences that are expected to occur in the future (e.g., a flight later in the day that the user is scheduled to take). Via a negative time offset, a suggested action can be provided to the user before the scheduled flight (e.g., requesting a carpool to the airport).

[0270] In one non-limiting embodiment, the anchor type associated with the user may be trained after a sufficient amount of training data has been acquired by the user interacting with the device upon wake-up. The training data may include signals generated by the device. A portion of these signals may encode multiple occurrences of one or more anchor types, and another portion of these signals may encode actions initiated by the user. The anchor occurrences encoded in the training data may be referred to as training anchor occurrences, and the actions encoded in the training data may be referred to as training actions. The user may initiate at least a portion of the training action in response to the training anchor occurrence. These signals may also encode metadata indicating at least some of the contextual conditions of the anchor occurrence. Each of the training actions may be classified as one or more action types of the action types enabled via the electronic device. More specifically, each training action may be classified as one of the action types in the set of enabled action types.

[0271] In some embodiments, anchor models may be trained for multiple anchor types in parallel. In other embodiments, models may be trained for anchor types in serial progression. After a sufficient amount of training data associated with a specific anchor type is collected, the filtering phase of the specific anchor type may be started. In various embodiments, a sufficient amount of training data for a specific anchor type may include the amount of training data encoding a statistically significant number of anchor occurrences of a specific anchor type and a statistically significant number of actions occurring simultaneously of one or more action types. As used herein, a statistically significant number of anchor occurrences and / or actions may refer to a sufficient amount of anchor occurrences and / or actions so that one or more statistical metrics (e.g., error bars, confidence levels, etc.) can be quantified within one or more thresholds. Note that when multiple actions are associated with a single anchor type, where the multiple actions are distinguished via the contextual conditions of the anchor occurrences, a larger amount of training data may be required.

[0272] During the filtering phase, various candidate action types are statistically correlated with the specific anchor type being trained based on the anchor occurrences encoded in the training data and the actions initiated by the user (e.g., the co-occurrence of anchors and actions in the training data). More specifically, a large, strong or high correlation between an anchor type and a specific action type indicates that the anchor and the action classified as a specific action type frequently occur simultaneously in the training data. A small, low or weak correlation between an anchor type and a specific action type indicates that the anchor and the action classified as a specific action type do not frequently occur simultaneously in the training data or there is no simultaneous occurrence. As discussed throughout, simultaneous occurrence in the training data does not need to be simultaneous in time, but can be separated by a time range that is characterized via one or more time offsets. The correlation can be conditioned by the contextual conditions or features of the anchor occurrence. The correlation is used to filter the enabled action type set of the electronic device to generate or determine the candidate action type set. Various methods can be used to determine the correlation between the anchor type and the action type. For example, various supervised or unsupervised machine learning (ML) methods can be used to identify patterns in the signal of the training data (e.g., the simultaneous occurrence of anchors and actions).

[0273] In the following discussion, anchor type training is discussed using an exemplary but non-limiting awakening anchor type. However, this discussion is non-limiting and is applicable to any number of other anchor types. Once a sufficient amount of training data associated with the awakening anchor type encodes a statistically significant amount of user awakening and a statistically significant amount of actions initiated when the user awakens, a filtering phase for the awakening anchor type training model can be performed. The filtering phase includes generating, determining and / or identifying a set of candidate action types that are subsets of the enabled action set. In various embodiments, the correlation between the anchor occurrence and the action type is determined via the training data. More specifically, the statistical correlation between the anchor type (e.g., awakening) and each enabled action type can be determined via the training data. Such statistical correlations between the anchor type and a specific action type may indicate the frequency of simultaneous occurrence of the anchor and the user initiating a specific action type. The correlation may be determined via one or more probability distributions that may be extracted from the training data. For example, one or more class-conditional probability distributions and / or one or more posterior probability distributions may be extracted from the training data. The probability distribution may be a condition for the simultaneous occurrence of the anchor and one or more actions. The correlation may be determined via the one or more probability distributions.

[0274] It should be noted that the simultaneous occurrence of anchor and action need not be simultaneous, but can be separated by a series of time values. As discussed below, the third learning stage (i.e., time offset stage) of training is transferred to determine the time offset based on the time value range between the occurrence of anchor and the initiation and / or execution of action. For example, if the user often plays a specific podcast shortly after waking up (e.g., usually between 10 minutes and 20 minutes), a relatively large or strong correlation between the anchor and the action type (e.g., playing a specific podcast) is determined. If the user rarely (or never) checks the bus schedule shortly after waking up, a relatively small or weak (or zero) correlation between the anchor and the action type of starting the bus schedule application is determined. The enabled action type (e.g., playing a podcast) that is more strongly related to the anchor type is included in the candidate action type set. The enabled action type (e.g., starting the bus schedule application) that is less related to the anchor type is not included in the candidate action type set. In some embodiments, the enabled action type set is filtered to determine the candidate action type set. The enabled action type of the "pass" filter is included in the candidate action type set.

[0275] As described above, various supervised or unsupervised ML methods can be used to determine the correlation. Generally speaking, ML methods can be used to identify the pattern of co-occurrence of anchors and events in training data associated with anchor types. The correlation used to identify the set of candidate action types can be determined based on the pattern of anchors and events in the training data (e.g., co-occurrence). A non-limiting example of an unsupervised ML method suitable for various embodiments is associated rule learning. In some embodiments, correlation can be generated by applying one or more associated rule learning methods and / or algorithms (such as but not limited to a priori learning algorithms) to the training data. Generally speaking, an associated rule learning method is used to analyze the co-occurrence of anchors and action types in the training data. The learned associated rules from co-occurring events (e.g., anchors and actions) are transferred to a priori rule learning as a type of algorithm that can be applied to training data to determine the correlation between anchor occurrence and action type.

[0276] In other embodiments, a neural network (NN) architecture that implements an attention mechanism may be used to determine the correlation between anchor occurrences and action types within training data. When implemented within a NN, the attention mechanism increases the ability of the NN to discover and / or learn patterns within the training data. Generally speaking, the analysis implemented by the NN involves parts of the training data where these patterns are more likely to be discovered. That is, the "attention" of the NN is shifted to a more appropriate part of the training data, rather than applying the same amount of attention to all parts of the training data. For example, with attention to "long-term patterns", patterns associated with "long-term" dependencies within the training data can be more easily discovered than NNs that do not implement an attention mechanism. Therefore, when an attention mechanism is used, it is easier to discover the simultaneous occurrence of anchors and actions. In other embodiments, other methods of analyzing patterns within the data may be used to determine the correlation between anchor occurrences and user actions. In other embodiments, a SparseMax algorithm may be used to apply to training data. That is, a deep learning NN may be used to identify patterns in the training data, wherein the activation function of at least one of the layers of the NN is a SparseMax function. The SparseMax activation function has the general property that the output vector is a relatively sparse vector. Therefore, the strongest correlation is identified in the training data. In some implementations, standard correlation analysis (CCA) may be employed to determine correlations in the training data.

[0277] Filtering the enabled action type set to generate the candidate action type set may be based on a threshold value applied to the correlation determined for each action type. In other embodiments, filtering may be based on the relative number and / or absolute number of action types to be included in the candidate action type set. In a non-limiting example, the filtering stage of training a model for a wake-up anchor type may filter the enabled action type set so that the candidate action type set includes five action types: play a podcast, play a song in the user's music library, view recent photos via a photo application, check the weather forecast via a weather application, or play a game installed on the device. When the candidate action type set is identified via the filtering stage, the sorting stage of the training cycle may begin.

[0278] During the sorting phase, the candidate action type set can be sorted, and at least a portion of the candidate action types can be selected based on the sorting. The sorting and selection of candidate action types can be based on contextual conditions or contextual features associated with the anchor occurrence encoded in the training data. That is, in response to (or expected) the occurrence of the anchor, the action types in the candidate action type set that the user is most likely to initiate or perform are highly sorted and selected for the sorting phase. The selected action type can be conditioned on the specific contextual conditions or contextual features of the anchor occurrence. In some embodiments, the sorting of specific action types of the candidate action type set can be determined based on training actions classified as specific action types. The sorting of specific action types can be further based on contextual conditions indicated by metadata associated with the training anchor occurrence, and the training anchor occurrence is related to the training action classified as a specific action type.

[0279] In the sorting stage, a decision tree learning method can be used to learn to distinguish between the contextual conditions of the anchor occurrence, and determine the recommended action based on the contextual conditions. For example, a classification decision tree algorithm can be used, such as but not limited to the Iterative Dichotomy 3 (ID3) algorithm can be applied to the relevant part of the training data to generate a classification decision tree for each action type in the candidate action type set (or each action type in the highest-level action type). The classification decision tree for a specific action type indicates which contextual conditions of the anchor occurrence are associated with the action type, and which contextual conditions are not associated with the anchor occurrence. In this way, the anchor model is trained to distinguish between the contextual conditions of the anchor occurrence, and the one or more recommended actions are customized for the anchor occurrence based on the contextual conditions. These nodes indicate one or more contextual conditions and whether the one or more contextual conditions "classify" the anchor occurrence (together with the contextual conditions) as being associated with the action type.

[0280] Fig. 8A A classification decision tree 800 is shown for the action type of playing a specific podcast and the anchor type of the user waking up according to various examples. The classification decision tree 800 is generated by applying a classification decision tree algorithm to the training data. It should be understood that the classification decision tree 800 is only an exemplary embodiment and the implementation scheme may vary. The classification decision tree typically subdivides the training data into smaller and smaller subsets of the training data. Each subset is conditioned on the contextual conditions associated with the node. Each node indicates whether the node (and therefore the associated subset of the training data) is based on the contextual conditions encountered when traversing the tree. The upper node 802 indicates that the classification decision tree 800 is used for the classification of "playing a specific podcast". The tree 800 is traversed (starting from the root node 804) by traversing the contextual conditions associated with the anchor occurrence. In this discussion, node 804 is called the root node because it indicates the first "decision" in the tree.

[0281] In a classification tree (e.g., tree 800), a root node (e.g., node 804) is typically conditioned on the contextual condition that most strongly predicts a class (e.g., playing a particular podcast). In tree 800, root node 804 is conditioned on whether an anchor (e.g., wakeup) occurs on a weekday or on a weekend. The root node indicates that the training data includes 44 samples where the anchor occurs, and 23 of the 44 samples occur on weekdays, and the remaining 13 samples occur on weekends (e.g., value = [31, 13]). The nodes in the decision tree may be characterized by a Gini metric, a Gini coefficient, a Gini index, which indicates an inequality of the underlying distribution, such as being conditioned on the contextual condition of the node. In some embodiments, the Gini metric may vary between 0.0 and 1.0 (or some other normalized range). Root node 804 is associated with a Gini metric of 0.499. If all anchors occur on weekdays, the Gini measure is likely to be closer to 1.0, whereas if all anchors occur on weekends, the Gini measure will be closer to 0.0.

[0282] Starting from the root node 804, if the anchor condition occurs on a weekday, the traversal process is to node 806. If the anchor occurs on the weekend, the traversal process is to node 808. Node 806 is conditioned on the contextual condition of playing a podcast at least 8 times in the first 10 days of the anchor occurrence. Node 808 is conditioned on the contextual condition that the anchor occurs on Saturday instead of Sunday. Nodes 806 and 808 provide sample numbers associated with distributions, values ​​of sample numbers, and Gini metrics. Nodes 810, 812, and 814 are leaf nodes and indicate various distributions, sample numbers of distributions, and associated Gini metrics. The class indicator at each node indicates whether the anchor occurrence with the associated contextual condition should be classified as a class (e.g., playing a specific podcast). Tree 800 indicates that when the user wakes up on a weekday and if the user has played podcasts frequently in the past 10 days (e.g., more than 7 times in the past 10 days), the action type of playing a specific podcast should be suggested.

[0283] Figure 8B A classification decision tree 820 is shown for the action type of playing a specific song and the anchor type of the user waking up according to various examples. The tree 820 is similar to Fig. 8A800, except that tree 820 is generated for the action type of playing a specific song from the user's music library when waking up. Tree 820 indicates the recommended action of playing the specific song when the user has listened to the song more than 6 times (when waking up) in the past 10 days but less than 23 times in the past 30 days. Therefore, when the user has listened to the song relatively frequently in the past 10 days and the user has not listened to the song too frequently in the past 30 days, the implementation scheme can provide a recommended action of playing the specific song. The classification decision tree algorithm can be used to generate such a classification decision tree for each action type in the candidate action type set. In some embodiments, as indicated by tree 800 and tree 820, if the contextual conditions of the anchor occurrence make the anchor occurrence categorized as both playing a specific podcast and playing a specific song, two action types can be suggested to the user. It should be noted that each action type can have a separate time offset. Therefore, a recommended action type can be provided at a separate time offset. For example, if the time offset for playing a podcast is 30 minutes and the time offset for playing a particular song is 15 minutes, an implementation may suggest playing the song for 15 minutes and waking up and suggest playing the podcast 30 minutes after waking up.

[0284] As described above, the time offset stage of training is transferred to the time offset of determining the suggested action type. In some embodiments, one or more candidate action types among the candidate action types may be selected for determining the time offset of the candidate action types. In some embodiments, the action type may be selected based on the ranking of the candidate action types and / or the classification decision tree generated for the candidate action types. In response to the anchor occurrence, the portion of the training data associated with the selected action type is used to generate a time distribution of the time when the user initiates the selected action type. The time distribution of the action type may be used to determine the time offset of the action type. As described throughout, the time offset of the action type may be used to determine the occurrence time point (after or before) of the anchor that provides the suggested action to the user.

[0285] In various embodiments, a histogram of the time distribution generated for the selected candidate action type may be generated. The time offset may be determined via the histogram. Fig. 9A bar graph 900 is shown of the temporal distribution of action types for playing a particular podcast in combination with wakeup anchors according to various examples. Bar graph 900 is provided as an example only, and embodiments may vary in the construction of the bar graphs for various action types. Bar graph 900 is generated from training data associated with anchor type wakeups and training actions initiated by a user in response to an anchor occurrence (e.g., a user initiating the playing of a podcast). The x-axis indicates the time (in seconds) between the anchor occurrence and the user initiating the playing of a particular podcast. In various embodiments, the x-axis is subdivided into multiple intervals. The y-axis indicates the number of times (within the training data) that the user initiated the action type within the time indicated by the x-axis value of the interval. Therefore, the y-axis value of each interval indicates the temporal distribution of the action type and the anchor occurrence. The time offset can be determined via a statistical measure of the time distribution. For example, in some non-limiting embodiments, the time offset for playing a particular podcast can be determined via at least one of the mean or median of the time distribution shown in bar graph 900. As Fig. 9 As shown, most of the time when users initiate playing a podcast upon waking up is within 700 seconds. Therefore, in this non-limiting example, the time offset of the action type of playing a specific podcast upon waking up can be set to approximately 10 minutes.

[0286] Although not shown, another time distribution (and histogram) may be generated for the action type of playing a particular song, and a separate time offset may be determined. In some embodiments, if sufficient statistics are included in the training data, a separate time distribution and / or histogram may be generated for each contextual condition in the anchor type contextual condition. For example, a separate histogram may be generated for anchor occurrences where the user wakes up on weekdays and anchor occurrences where the user wakes up on weekends.

[0287] When the filtering, sorting, and time shifting phases of training are completed, the anchor model can be updated to provide suggested actions, as learned via training. Once updated, the anchor model can be deployed to the device to start providing suggested actions in response to (or in anticipation of) anchor occurrences. In various embodiments, once deployed at an electronic device, the model can be iteratively updated periodically or from time to time based on user feedback. That is, the model can be iteratively updated based on the user's participation in the suggested action. When the user actively participates in a specific suggested action (e.g., the user selects a specific suggested action to initiate its execution), the model is updated to enhance or increase the probability that the model provides a suggested action in response to future anchor occurrences. When the user does not participate (or passively participates) in a specific suggested action (e.g., the user does not select a specific suggested action to initiate its execution), the model is updated to reduce the probability that the model provides a specific suggested action in response to future anchor occurrences. In some embodiments, one or more reinforcement learning methods can be used to iteratively update the anchor model based on the user's participation in the suggested action or other user feedback. When the user selects a suggested action, the reward function adopted in such reinforcement learning embodiments may be greater than when the user does not select a suggested action. Therefore, the anchor model can be iteratively updated based on the user's responses to the suggested actions.

[0288] Fig.10 A process 1000 for training a prediction model according to various examples that enables suggested actions to be provided in response to the occurrence of an anchor event is shown. For example, process 1000 is performed using one or more electronic devices that implement a digital assistant. In some examples, process 1000 is performed using a client-server system (e.g., system 100), and the boxes of process 1000 are divided in any manner between a server (e.g., DA server 106) and a client device. In other examples, the boxes of process 1000 are divided between a server and multiple client devices (e.g., a mobile phone and a smart watch). Therefore, although portions of process 1000 are described herein as being performed by specific devices of a client-server system, it should be understood that process 1000 is not limited to this. In other examples, process 1000 is performed using only a client device (e.g., user device 104) or only multiple client devices. In process 1000, some boxes are optionally combined, the order of some boxes is optionally changed, and some boxes are optionally omitted. In some examples, additional steps may be performed in conjunction with process 1000. The above reference Fig.10 The described operations are optionally performed by Figures 1 to 4 , FIG. 6A to FIG. 6B and 7A to 7C A person skilled in the art will clearly know how to implement the Figures 1 to 4 , FIG. 6A to FIG. 6B and 7A to 7C The components depicted in the figure are used to implement other processes.

[0289] Process 1000 begins at box 1002, where training data is obtained for training a prediction model, such as but not limited to an anchor model. Once trained, the anchor model is used to provide suggested actions in response to (or in anticipation of) the occurrence of an anchor and / or an anchor event. An anchor can be an event indicating one or more user behaviors of an electronic device. Therefore, in the foregoing discussion, an anchor can be referred to as an event, and the occurrence of an anchor can be referred to as an event occurrence. The training data includes a signal generated by an electronic device that enables an enabled action set. The signal can be referred to as a training signal. At box 1004, the signal can be analyzed to identify an event occurrence, an associated contextual condition, and an action initiated by a user of the device in response to (or in anticipation of) the event occurrence. The event occurrence, the associated contextual condition, and the training action can be encoded in the signal. The contextual condition can be included in the metadata associated with the event occurrence. Each training action in the training action can be classified as one or more action types in the action type of the enabled action set.

[0290] At box 1006, the statistical correlation between the event occurrence and the action type based on the training signal. In some embodiments, the correlation between the event occurrence and at least a portion of the training action can be determined. The statistical correlation can be determined via any one of the various embodiments discussed herein. For example, the correlation can be determined via one or more supervised or unsupervised machine learning (ML) methods and / or algorithms. At box 1008, a candidate action type set can be determined based on the correlation between the event occurrence and the action type. The candidate action type set can be a subset of the enabled action set. In some embodiments, the enabled action type set can be filtered to generate and / or identify the candidate action set. Any one of the various filters and / or filtering methods discussed herein can be used to filter the enabled action type set. In some embodiments, boxes 1006 and 1008 may include a filtering stage for training a prediction model.

[0291] At box 1010, a ranking for each action type in the action type of the candidate action type set may be determined and / or generated. The ranking may be based on a portion of the training action that is classified as the ranked action type, a contextual condition of an event occurrence associated with the portion of the training action, and a correlation between the portion of the training action and the associated event occurrence. At box 1012, one or more action types may be selected from the candidate action type set. The selection of action types may be based on the ranking of the candidate action types. In various embodiments, boxes 1010 and 1012 may include a ranking phase of training a prediction model. Therefore, at boxes 1010 and / or 1012, one or more classification decision trees may be generated for each ranked action type in the ranked action types, as discussed throughout. The classification decision trees may be based on the contextual condition of an associated event occurrence. The selection of action types may be based on the classification decision trees. In combination with at least FIG. 8A to FIG. 8B Let's discuss non-limiting examples of such classification decision trees.

[0292] At block 1014, one or more time distributions may be generated for each action type selected at block 1012. The time distribution for a particular action type may indicate a time difference between an event occurrence and an associated training action classified as the selected action type. A separate time distribution may be generated for each action type selected at block 1012. A separate time distribution may be generated for each of the different contextual conditions under which the event occurs. A histogram may be generated for each of the time distributions. In combination with at least Fig. 9 A non-limiting example of a histogram is discussed. At block 1016, a time offset is determined for each of the one or more action types selected at block 1012. The time offset may be determined based on a time distribution and / or histogram generated for the selected action type. In various embodiments, blocks 1014 and 1016 may include a time offset phase for training a prediction model.

[0293] At block 1018, the prediction model may be updated based on the training of process 1000. In some embodiments, the prediction model may be updated to provide one or more suggested actions to the user at the electronic device and in response to the occurrence of another event, based on the event type of the other event, one or more of the selected action types, the one or more time offsets of the selected one or more action types, and the contextual conditions of the other event occurrence. That is, the trained event or anchor model may be deployed at the electronic device. In combination with at least Fig.11 Various embodiments are described for deploying trained predictive models on electronic devices.

[0294] Fig.11 A process 1100 for deploying a trained prediction model at an electronic device according to various examples is shown. Similar to process 1000, process 1100 can be performed, for example, using one or more electronic devices implementing a digital assistant. Process 11100 can be performed by any device and / or combination of devices similar to the device and / or combination of devices by which process 1000 can be performed.

[0295] Process 1100 begins at block 1102 where an event (or anchor) occurrence and contextual conditions are detected via monitoring and analyzing signals generated by an electronic device. At block 1104, one or more suggested actions and one or more time offsets are determined based on the detected event occurrence and associated contextual conditions using a trained prediction model. In some embodiments, the one or more suggested actions and the one or more associated time offsets are received at the electronic device via the prediction model. In some embodiments, the prediction model may be composed of at least one combination of Fig.10 At block 1106, the electronic device is employed to provide the one or more suggested actions to the user according to the associated one or more time offsets. At block 1108, in response to receiving an indication and / or selection of the provided suggested action, execution of the selected suggested action is initiated via the electronic device.

[0296] In various embodiments, once deployed at an electronic device, the model may be iteratively updated periodically or from time to time based on user feedback. That is, the model may be iteratively updated based on user engagement with the suggested actions. Such iterative updates are not Fig.11 . However, it should be understood that process 1100 can be modified to include iteratively updating the prediction model based on user participation and / or feedback in response to the suggested action. When the user actively participates in a specific suggested action (e.g., the user selects a specific suggested action to initiate its execution), the model is updated to enhance or increase the probability that the model provides a suggested action in response to the occurrence of a future anchor. When the user does not participate (or passively participates) in a specific suggested action (e.g., the user does not select a specific suggested action to initiate its execution), the model is updated to reduce the probability that the model provides a specific suggested action in response to the occurrence of a future anchor. In some embodiments, one or more reinforcement learning methods can be used to iteratively update the anchor model based on the user's participation in the suggested action or other user feedback. When the user selects a suggested action, the reward function used in such reinforcement learning implementation schemes may be greater than when the user does not select a suggested action. Therefore, the anchor model can be iteratively updated based on the user's response to the suggested action.

[0297] According to some specific implementations, a computer-readable storage medium (e.g., a non-transitory computer-readable storage medium) is provided, which stores one or more programs for execution by one or more processors of an electronic device, the one or more programs including instructions for performing any of the methods or processes described herein.

[0298] According to some implementations, an electronic device (eg, a portable electronic device) is provided that includes means for performing any of the methods and processes described herein.

[0299] According to some implementations, an electronic device (eg, a portable electronic device) is provided that includes a processing unit configured to perform any of the methods and processes described herein.

[0300] According to some specific implementations, an electronic device (e.g., a portable electronic device) is provided, which includes one or more processors and a memory storing one or more programs for execution by the one or more processors, the one or more programs including instructions for performing any of the methods and processes described herein.

[0301] Example methods are disclosed herein. An example method may be used to train a prediction model (e.g., and anchor model) for an anchor (or event) that indicates a user behavior of an electronic device that can perform each action type in an enabled action type set. The electronic device may have one or more processors and a memory. The electronic device may perform one or more actions and / or operations. The method may include multiple training actions and multiple event occurrences of events detected at an electronic device and based on multiple signals generated by the electronic device. Each event in the multiple event occurrences may be associated with metadata indicating contextual conditions of the event occurrence. Each training action in the multiple training actions may be initiated via user interaction with the electronic device, and each training action in the multiple training actions may be classified as an action type in the enabled action type set. A candidate action type set may be determined. Determining the candidate action type set may be based on multiple correlations between each training action in the multiple training actions and each event occurrence in the multiple event occurrences. The candidate action type set may be a subset of the enabled action type set. The ranking of each action type in the candidate action type set may be determined. Determining the ranking of action types may be based on a portion of the multiple training actions classified as action types. Determine the order of action types can be further based on the one or more contextual conditions indicated by the metadata associated with a portion of the multiple events, and a portion of the multiple events is related to the portion of the multiple training actions via the multiple correlations. In some embodiments, the first action type of the candidate action type set can be selected based on the order determined for each action type of the candidate action type set. The first part of the multiple training actions and the first part of the multiple events can be selected. Each part in the first part of the multiple training actions can be classified as a first action type. Each part in the first part of the multiple events can be related to at least one of the first parts of the multiple training actions via the multiple correlations. The time offset of the first action type can be determined based on the time distribution of the first part of the multiple events relative to the first part of the multiple events. In some embodiments, the prediction model can be updated to generate a suggested action in response to another occurrence of the event. The suggested action provided is based on the time offset of the first action type and the first action type.

[0302] The example method may include using an association rule learning algorithm to determine the multiple correlations between each of the multiple training actions and each of the multiple event occurrences. In some embodiments, a standard correlation analysis may be used to determine the multiple correlations between each of the multiple training actions and each of the multiple event occurrences. In other embodiments, attention within a neural network may be used to determine the multiple correlations between each of the multiple training actions and each of the multiple event occurrences. In one or more other embodiments, a SparseMax algorithm may be used to determine the multiple correlations between each of the multiple training actions and each of the multiple event occurrences.

[0303] Determining the candidate action type set may include applying a filter to the enabled action type set, and the filter is based on the multiple correlations between each training action in the multiple training actions and each event occurrence in the multiple event occurrences. The filter may be an entropy filter. In other embodiments, the filter is a classical conditional probability filter. In other embodiments, the filter is one of a local maximum filter or a local minimum filter. In some embodiments, the filter may include filtering the number of times the candidate action type has been initiated in response to (or expectation) the occurrence of the anchor in the previous amount of time (e.g., the number of times the user has initiated the candidate action in the previous 28 days). In other embodiments, the filter may filter the posterior probability associated with the candidate action type. In other embodiments, the filter may filter the class conditional probability associated with the candidate action type (e.g., the probability that the anchor is conditioned on the candidate action type). The filter may be a combination of multiple filters. For example, the filter may be a combination of filters about the number of times (or frequency) the user has initiated the candidate action type in response to the anchor, filters about the posterior probability of the candidate action type, and filters about the class conditional probability of the anchor conditioned on the candidate action type.

[0304] This example method may also include generating a classification decision tree for each action type of the candidate action type set. The nodes of the classification decision tree can be conditional on the contextual conditions associated with the parts of the multiple training actions that are classified as action types that occur with these multiple events. The classification decision tree can be classified as being associated with action types with contextual conditions by other events. In some embodiments, the first action type of the candidate action type set can be further based on the classification decision tree of the first action type. The classification decision tree of each action type of the candidate action type set can be used to associate multiple action types of the candidate action type set with events. These multiple action types can be distinguished by the contextual conditions of the multiple events that occur.

[0305] The example method may also include generating a first histogram for a first action type based on a time distribution of a first training action in the plurality of training actions. A time offset of the first action type may be determined based on the first histogram. The method may also include generating a second histogram for the first action type based on a second time distribution of the plurality of training actions. The time distribution may be conditioned on a first set of contextual conditions, and the second time distribution may be conditioned on a second set of contextual conditions. A second time offset of the first action type may be determined based on the second time distribution of the plurality of training actions. The prediction model may be updated to generate a suggested action based on the first action type and the time offset of the first action type in response to a first occurrence of an event conditioned on the first set of contextual conditions. The prediction model may be further updated to generate a second suggested action based on the first action type and the second time offset of the first action type in response to a second occurrence of an event conditioned on the second set of contextual conditions.

[0306] The time offset of the first action type may be further based on at least one of a mean or a median of a time distribution of the plurality of training actions. The prediction model may be stored and / or encrypted on the electronic device. The time offset may be less than zero or greater than zero. Generating the suggested action may include providing a notification at a time when the electronic device is at a time that is separated from other events by a time offset.

[0307] Another example method may be used to employ a predictive model for an event indicative of user behavior of an electronic device. Other methods may include detecting an event occurrence of the event based on one or more signals generated by the electronic device. The event occurrence may be associated with metadata indicating contextual conditions of the event occurrence. Based on the event occurrence, a suggested action and a time offset may be received from the predictive model. The suggested action may be provided to the user within a time offset relative to the event occurrence.

[0308] In the method, the event may include the termination of the do not disturb mode of the electronic device, and the user behavior includes user waking up. Detecting the occurrence of the event may include determining that the electronic device is located at a specific location based on a positioning signal in the one or more signals. The specific location may include at least one of an airport, a bus terminal, or a train terminal. In other embodiments, the specific location includes at least one of a user's residence, a user's workplace, an entertainment location, or a fitness center. The occurrence of the event may include the occurrence of a calendar event of an electronic calendar associated with the user. Providing a suggested action may include providing a suggested action after the event occurs and according to a time offset. In other embodiments, providing a suggested action includes providing a suggested action before the event occurs and according to a time offset. In other embodiments, providing a suggested action includes using a display of an electronic device to provide a pop-up notification indicating a suggested action.

[0309] The method may also include receiving a user selection of a suggested action at an electronic device. The electronic device may be used to automatically perform the suggested action. The prediction model may be trained based on a plurality of training signals generated by the electronic device. One or more machine learning (ML) methods may be used to train the prediction model via the training signal. In some embodiments, the method also includes providing a second suggested action to the user within a second time offset relative to the occurrence of the event according to the prediction model. The suggested action may be based at least on the event type of the event occurrence. The suggested action may be further based on at least one or more contextual conditions of the event occurrence. The one or more contextual conditions of the event occurrence may include at least one of a day of the week, a time of the day, or the location of the electronic device. The suggested action may include at least one of playing content on the electronic device, launching an application installed on the electronic device, or sending a message to a first contact in a plurality of contacts of the user. The application installed on the electronic device may include at least one of a carpooling application, a physical fitness application, or a food delivery application.

[0310] Some exemplary embodiments of the present application are shown below.

[0311] Embodiment 1. A method for training a prediction model for an event, the event indicating a user behavior of an electronic device, the electronic device being enabled to perform each action type in a set of enabled action types, the method comprising:

[0312] One or more processors and memory of the electronic device are used to perform operations including:

[0313] detecting, based on a plurality of signals generated by the electronic device, a plurality of training actions and a plurality of event occurrences of the event, wherein each of the plurality of event occurrences is associated with metadata indicating a contextual condition of the event occurrence, and each of the plurality of training actions is initiated via interaction between the user and the electronic device and is classified as an action type in the set of enabled action types;

[0314] determining a set of candidate action types based on a plurality of correlations between each of the plurality of training actions and each of the plurality of event occurrences, wherein the set of candidate action types is a subset of the enabled set of action types;

[0315] determining a ranking for each action type of the set of candidate action types based on a portion of the plurality of training actions classified as the action type and the one or more contextual conditions indicated by the metadata associated with a portion of the plurality of event occurrences, the portion of the plurality of event occurrences being related to the portion of the plurality of training actions via the plurality of correlations;

[0316] selecting a first action type in the set of candidate action types based on the determined ranking for each action type in the set of candidate action types;

[0317] selecting a first portion of the plurality of training movements and a first portion of the plurality of event occurrences, wherein each portion of the first portion of the plurality of training movements is classified as the first movement type and each portion of the first portion of the plurality of event occurrences is correlated with at least one of the first portion of the plurality of training movements via the plurality of correlations;

[0318] determining a time offset of the first action type based on a time distribution of the plurality of training actions relative to the first portion of the plurality of event occurrences; and updating the prediction model to generate a suggested action based on the first action type and the time offset of the first action type in response to another occurrence of the event.

[0319] Embodiment 2. The method according to embodiment 1 further includes:

[0320] An association rule learning algorithm is employed to determine the plurality of correlations between each of the plurality of training actions and each of the plurality of event occurrences.

[0321] Embodiment 3. The method according to any one of embodiments 1 to 2, further comprising:

[0322] A standard correlation analysis is employed to determine the plurality of correlations between each training move in the plurality of training moves and each event occurrence in the plurality of event occurrences.

[0323] Embodiment 4. The method according to any one of embodiments 1 to 3, further comprising:

[0324] Attention is employed within a neural network to determine the plurality of correlations between each of the plurality of training actions and each of the plurality of event occurrences.

[0325] Embodiment 5. The method according to any one of embodiments 1 to 4, further comprising:

[0326] A SparseMax algorithm is employed to determine the plurality of correlations between each of the plurality of training actions and each of the plurality of event occurrences.

[0327] Embodiment 6. The method according to any one of embodiments 1 to 5, wherein the determination

[0328] The candidate action type set includes applying a filter to the enabled action type set, and the filter is based on the plurality of correlations between each of the plurality of training actions and each of the plurality of event occurrences.

[0329] Embodiment 7. A method according to embodiment 6, wherein the filter is an entropy filter.

[0330] Embodiment 8. A method according to any one of embodiments 6 to 7, wherein the filter is a class conditional probability filter.

[0331] Embodiment 9. The method of any one of Embodiments 6 to 8, wherein the filter is one of a local maximum filter or a local minimum filter.

[0332] Embodiment 10. The method according to any one of embodiments 1 to 9, further comprising:

[0333] For each action type in the set of candidate action types, a classification decision tree is generated, wherein the nodes of the classification decision tree are conditioned on contextual conditions associated with the portion of the multiple event occurrences related to the portion of the multiple training actions classified as the action type, and the classification decision tree uses the contextual conditions to classify other event occurrences as associated with the action type.

[0334] Embodiment 11. The method according to embodiment 10, wherein selecting the first action type of the set of candidate action types is further based on the classification decision tree of the first action type.

[0335] Embodiment 12. The method according to any one of embodiments 10 to 11, further comprising:

[0336] The classification decision tree is used for each action type in the candidate action type set to associate multiple action types in the candidate action type set with the event, wherein the multiple action types are distinguished by the context conditions under which the multiple events occur.

[0337] Embodiment 13. The method according to any one of embodiments 1 to 12, further comprising:

[0338] generating a first histogram for the first movement type based on the time distribution of a first training movement of the plurality of training movements; and

[0339] The time offset of the first action type is determined based on the first histogram.

[0340] Embodiment 14. The method according to any one of embodiments 1 to 13, further comprising:

[0341] generating a second histogram for the first action type based on a second temporal distribution of the plurality of training actions, wherein the temporal distribution is conditioned on a first set of contextual conditions and the second temporal distribution is conditioned on a second set of contextual conditions;

[0342] determining a second time offset of the first action type based on the second time distribution of the plurality of training actions;

[0343] updating the prediction model to generate the suggested action based on the first action type and the time offset of the first action type in response to a first occurrence of the event conditioned on the first set of contextual conditions; and

[0344] The prediction model is updated to generate a second suggested action based on the first action type and the second time offset of the first action type in response to a second occurrence of the event conditioned on the second set of contextual conditions.

[0345] Embodiment 15. A method according to any one of embodiments 1 to 14, wherein the time offset of the first action type is further based on at least one of a mean or a median of the time distribution of the plurality of training actions.

[0346] Embodiment 16. A method according to any one of embodiments 1 to 15, wherein the prediction model is stored on the electronic device.

[0347] Embodiment 17. A method according to any one of embodiments 1 to 16, wherein the time offset is less than zero.

[0348] Embodiment 18. A method according to any one of embodiments 1 to 17, wherein generating the suggested action includes providing a notification at the electronic device at a time separated from the occurrence of the other event by the time offset.

[0349] Embodiment 19. An electronic device comprising:

[0350] one or more processors;

[0351] Memory;

[0352] One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs include instructions for training a prediction model for an event, the event indicating a user behavior of the electronic device, the electronic device being enabled to perform each action type in the enabled action type set, and the instructions include operations of:

[0353] detecting, based on a plurality of signals generated by the electronic device, a plurality of training actions and a plurality of event occurrences of the event, wherein each of the plurality of event occurrences is associated with metadata indicating a contextual condition of the event occurrence, and each of the plurality of training actions is initiated via interaction between the user and the electronic device and is classified as an action type in the set of enabled action types;

[0354] determining a set of candidate action types based on a plurality of correlations between each of the plurality of training actions and each of the plurality of event occurrences, wherein the set of candidate action types is a subset of the enabled set of action types;

[0355] determining a ranking for each action type of the set of candidate action types based on a portion of the plurality of training actions classified as the action type and the one or more contextual conditions indicated by the metadata associated with a portion of the plurality of event occurrences, the portion of the plurality of event occurrences being related to the portion of the plurality of training actions via the plurality of correlations;

[0356] selecting a first action type in the set of candidate action types based on the determined ranking for each action type in the set of candidate action types;

[0357] selecting a first portion of the plurality of training movements and a first portion of the plurality of event occurrences, wherein each portion of the first portion of the plurality of training movements is classified as the first movement type and each portion of the first portion of the plurality of event occurrences is correlated with at least one of the first portion of the plurality of training movements via the plurality of correlations;

[0358] determining a time offset of the first action type based on a time distribution of the multiple training actions relative to the first portion of the multiple event occurrences; and updating the prediction model to generate a suggested action based on the first action type and the time offset of the first action type in response to another occurrence of the event.

[0359] Embodiment 20. A non-transitory computer-readable storage medium storing one or more programs, the one or more programs including instructions for training a prediction model for an event, the event indicating a user behavior of an electronic device, the electronic device being enabled to perform each action type in an enabled action type set, which when executed by one or more processors of the electronic device causes the electronic device to perform operations including:

[0360] detecting, based on a plurality of signals generated by the electronic device, a plurality of training actions and a plurality of event occurrences of the event, wherein each of the plurality of event occurrences is associated with metadata indicating a contextual condition of the event occurrence, and each of the plurality of training actions is initiated via interaction between the user and the electronic device and is classified as an action type in the set of enabled action types;

[0361] determining a set of candidate action types based on a plurality of correlations between each of the plurality of training actions and each of the plurality of event occurrences, wherein the set of candidate action types is a subset of the enabled set of action types;

[0362] determining a ranking for each action type of the set of candidate action types based on a portion of the plurality of training actions classified as the action type and the one or more contextual conditions indicated by the metadata associated with a portion of the plurality of event occurrences, the portion of the plurality of event occurrences being related to the portion of the plurality of training actions via the plurality of correlations;

[0363] selecting a first action type in the set of candidate action types based on the determined ranking for each action type in the set of candidate action types;

[0364] selecting a first portion of the plurality of training movements and a first portion of the plurality of event occurrences, wherein each portion of the first portion of the plurality of training movements is classified as the first movement type and each portion of the first portion of the plurality of event occurrences is correlated with at least one of the first portion of the plurality of training movements via the plurality of correlations;

[0365] determining a time offset of the first action type based on a temporal distribution of the plurality of training actions relative to the first portion of the plurality of event occurrences; and

[0366] The predictive model is updated to generate a suggested action based on the first action type and the time offset of the first action type in response to another occurrence of the event.

[0367] For the purpose of explanation, the preceding description is described by reference to specific embodiments. However, the above exemplary discussion is not intended to be exhaustive or limit the invention to the disclosed precise form. According to the above teachings, many modifications and variations are possible. These embodiments are selected and described in order to best explain the principles of these technologies and their practical applications. Others skilled in the art can thus best utilize these technologies and various embodiments with various modifications suitable for the intended specific use.

[0368] Although the disclosure and examples have been fully described with reference to the accompanying drawings, it should be noted that various changes and modifications will become apparent to those skilled in the art. It should be understood that such changes and modifications are considered to be included within the scope of the disclosure and examples defined by the claims.

[0369] As described above, one aspect of the present technology is to collect and use data available from various sources to provide suggested actions in response to detecting events and / or anchors. The present disclosure contemplates that, in some instances, these collected data may include personal information data that uniquely identifies or can be used to contact or locate a specific person. Such personal information data may include demographic data, location-based data, phone numbers, email addresses, Twitter IDs, home addresses, data or records related to the user's health or fitness level (e.g., vital sign measurements, medication information, exercise information), date of birth, or any other identifying or personal information.

[0370] The present disclosure recognizes that the use of such personal information data in the present technology can be used to benefit users. For example, personal information data can be used to provide suggested actions to users in response to detected events and / or anchors. Therefore, the use of such personal information data enables users to have planned control over the content delivered. In addition, the present disclosure also anticipates other uses of personal information data that benefit users. For example, health and fitness data can be used to provide insights into the user's overall health status, or can be used as positive feedback to individuals who use technology to pursue health goals.

[0371] This disclosure envisions that entities responsible for collecting, analyzing, disclosing, transmitting, storing or otherwise using such personal information data will comply with established privacy policies and / or privacy practices. Specifically, such entities should implement and adhere to privacy policies and practices that are recognized as meeting or exceeding industry or government requirements for maintaining the privacy and security of personal information data. Such policies should be easily accessible to users and should be updated as the collection and / or use of data changes. Personal information from users should be collected for legitimate and reasonable purposes of the entity and not shared or sold outside of these legitimate uses. In addition, such collection / sharing should be performed after receiving informed consent from the user. In addition, such entities should consider taking any necessary steps to defend and secure access to such personal information data and ensure that others who have access to personal information data comply with their privacy policies and processes. In addition, such entities may subject themselves to third-party assessments to demonstrate their compliance with widely accepted privacy policies and practices. In addition, policies and practices should be adjusted to specific types of personal information data collected and / or accessed, and to applicable laws and standards including specific considerations of jurisdiction. For example, in the United States, the collection or access of certain health data may be governed by federal and / or state laws, such as the Health Insurance Portability and Accountability Act (HIPAA), while health data in other countries may be subject to other regulations and policies and should be handled accordingly. Therefore, different privacy practices should be maintained in each country for different types of personal data.

[0372] Regardless of the foregoing, the present disclosure also contemplates implementation schemes in which users selectively block the use or access of personal information data. That is, the present disclosure contemplates that hardware elements and / or software elements may be provided to prevent or block access to such personal information data. For example, in the case of providing suggested actions, the technology of the present invention may be configured to allow users to choose to "opt in" or "opt out" to participate in the collection of personal information data during registration for a service or at any time thereafter. As another example, a user may choose not to provide training data and / or trained models to other devices and / or other users. In another example, a user may choose to limit the length of time that emotion-related data is retained, or completely prohibit the development of basic emotional conditions. In addition to providing "opt-in" and "opt-out" options, the present disclosure contemplates providing notifications related to access or use of personal information. For example, a user may be notified that their personal information data will be accessed when downloading an application, and then the user may be reminded again just before the personal information data is accessed by the application.

[0373] Furthermore, it is an objective of the present disclosure to manage and process personal information data to minimize the risk of unintentional or unauthorized access or use. Risks can be minimized by limiting data collection and deleting data once it is no longer needed. Additionally, and when applicable, including in certain health-related applications, data de-identification can be used to protect the privacy of users. Where appropriate, de-identification can be facilitated by removing specific identifiers (e.g., date of birth, etc.), controlling the amount or characteristics of stored data (e.g., collecting location data at a city level rather than an address level), controlling how data is stored (e.g., aggregating data across users), and / or other methods.

[0374] Thus, while the present disclosure broadly covers the use of personal information data to implement one or more of the various disclosed embodiments, the present disclosure also contemplates that the various embodiments may be implemented without access to such personal information data.

Claims

1. A method for employing a predictive model for an event, the event indicating a behavior of a user of an electronic device, the method comprising: One or more processors and memory of the electronic device are used to perform operations including: detecting an event occurrence of the event based on one or more signals generated by the electronic device, wherein the event occurrence is associated with metadata indicating contextual conditions of the event occurrence; receiving a suggested action and a time offset from the prediction model based on the occurrence of the event; as well as The suggested action is provided to the user within the time offset relative to an occurrence of the event. 2 . The method of claim 1 , wherein the event comprises termination of a do not disturb mode of the electronic device, and the user's behavior comprises waking up of the user. 3 . The method according to claim 1 , wherein detecting the occurrence of the event comprises determining that the electronic device is located at a specific location based on a positioning signal in the one or more signals. The method of claim 3 , wherein the specific location comprises at least one of an airport, a bus terminal, or a train terminal.

5. The method of claim 3, wherein the specific location comprises at least one of the user's residence, the user's workplace, an entertainment location, or a fitness center.

6. The method of any one of claims 1 to 2, wherein the event occurrence comprises an occurrence of a calendar event of an electronic calendar associated with the user.

7. The method of any one of claims 1 to 2, wherein providing the suggested action comprises providing the suggested action after the event occurs and according to the time offset.

8. The method of any one of claims 1 to 2, wherein providing the suggested action comprises providing the suggested action before the event occurs and according to the time offset.

9. The method of any one of claims 1 to 2, wherein providing the suggested action comprises providing a pop-up notification indicating the suggested action using a display of the electronic device.

10. The method according to any one of claims 1 to 2, further comprising: receiving, at the electronic device, a user selection of the suggested action; as well as The electronic device is employed to perform the suggested action.

11. The method according to any one of claims 1 to 2, wherein the prediction model is trained based on a plurality of training signals generated by the electronic device.

12. The method according to any one of claims 1 to 2, wherein the prediction model is trained using one or more machine learning (ML) methods.

13. The method according to any one of claims 1 to 2, further comprising: Based on the predictive model, a second suggested action is provided to the user within a second time offset relative to the occurrence of the event.

14. The method of any one of claims 1 to 2, wherein the suggested action is based at least on an event type of the event occurrence.

15. The method according to any one of claims 1 to 2, wherein the suggested action is based on at least one or more contextual conditions of the occurrence of the event.

16. The method of claim 15, wherein the one or more contextual conditions under which the event occurs include at least one of a day of the week, a time of the day, or a location of the electronic device.

17. The method according to any one of claims 1 to 2, wherein the suggested action comprises at least one of playing content on the electronic device, launching an application installed on the electronic device, or sending a message to a first contact among multiple contacts of the user.

18. The method of claim 17, wherein the application installed on the electronic device comprises at least one of a ride-sharing application, a fitness application, or a food delivery application.

19. An electronic device comprising: one or more processors; Memory; One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs include instructions for employing a predictive model for an event, the event indicating a behavior of a user of the electronic device, the electronic device being enabled to perform each action type in a set of enabled action types, and the instructions including operations according to any one of claims 1 to 18.

20. A computer program product comprising one or more programs, wherein the one or more programs include instructions for employing a predictive model for an event, wherein the event indicates the behavior of a user of an electronic device, and wherein the instructions, when executed by one or more processors of the electronic device, cause the electronic device to perform the operations of the method described in any one of claims 1 to 18.

Images