Information processing method and device and electronic equipment

Abstract

The invention discloses an information processing method and device and electronic equipment, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring a first audio signal, wherein the first audio signal comprises a human voice signal; displaying a first window under the condition that a schedule creation user intention is recognized based on the first audio signal; wherein the first window comprises first schedule information, first calendar information and first prompt information; the first schedule information comprises description information of schedule elements of a first schedule; the schedule content of the first schedule is determined based on the human voice signal; the first prompt information is used for prompting whether to create the first schedule; the first calendar information comprises first schedule time; the first schedule time is determined based on the human voice signal.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of artificial intelligence, and particularly relates to an information processing method and device and electronic equipment. BACKGROUND

[0002] In today's fast-paced information age, it is a hot topic to quickly capture key information needed by oneself from numerous fleeting information and quickly record the captured key information for subsequent information tracing.

[0003] Currently, there are problems of low schedule creation efficiency and high schedule creation error rate in quickly capturing and recording the schedule needed by the user from numerous fleeting information. For example, in a social occasion such as business negotiation or meeting where information exchange is fast, the key schedule needs to be quickly captured and recorded by the user. The usual practice is to use an electronic device to record the whole process of the meeting, but this way requires the user to spend a lot of time listening to the recording after the meeting, and then the user creates the schedule in his own schedule application, resulting in low schedule creation efficiency. The user can also simply record during the meeting, and then summarize and create the schedule in detail after the meeting, which is easy to cause the created schedule to be incorrect due to the user's fuzzy memory, and simply recording the schedule during the meeting will interrupt the task the user is currently doing. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide an information processing method, device, electronic equipment and storage medium, which can improve the accuracy and efficiency of schedule creation, and avoid the interruption of the task in hand of the user due to manual creation of the schedule in a serious immersive scenario such as business negotiation.

[0005] In a first aspect, the embodiments of the present application provide an information processing method, which comprises: obtaining a first audio signal, wherein the first audio signal comprises a human voice signal; displaying a first window in a case where a schedule creation user intention is recognized based on the first audio signal; wherein the first window comprises first schedule information, first calendar information and first prompt information; the first schedule information comprises description information of a schedule element of a first schedule; the schedule content of the first schedule is determined based on the human voice signal; the first prompt information is used to prompt whether to create the first schedule; and the first calendar information comprises a first schedule time; and the first schedule time is determined based on the human voice signal.

[0006] In this way, in a case where a schedule creation user intention is recognized in the acquired first audio signal containing a human voice signal, a first window can be displayed, the first window including first schedule information, first calendar information, and first prompt information, the first schedule information including description information of a schedule element of a first schedule, a schedule content of the first schedule being determined based on the human voice signal, the first prompt information being used to prompt whether to create the first schedule, and the first calendar information including a first schedule time; the first schedule time being determined based on the human voice signal. In this way, the electronic device can automatically and accurately extract core elements such as a time and an event in a schedule from continuous human voice signals and automatically fill them into the first window, without the need for the user to listen back to the full recording in the electronic device or for the user to stop the current task to create a schedule, thereby saving the user from a series of cumbersome operations such as manually opening a calendar application, inputting text one by one, and selecting a date, and only needing to confirm the first schedule information generated on the first window, thereby simplifying a plurality of operations into one confirmation, greatly shortening the path and time of schedule creation, improving the accuracy and efficiency of schedule creation, and avoiding interruption of the current task due to manual schedule creation in a serious immersive scenario such as a business meeting.

[0007] In some embodiments of the present application, after the first audio signal is acquired, the method further includes: performing semantic recognition on the first audio signal to obtain a first recognition result; the first recognition result is used to indicate whether a schedule creation user intention is recognized based on the first audio signal; in a case where the first recognition result indicates that a schedule creation user intention is recognized based on the first audio signal, cutting, from the first audio signal, an audio signal containing the schedule creation user intention to obtain a second audio signal; generating first schedule information according to the second audio signal.

[0008] In this way, in a case where a schedule creation user intention is recognized based on the first audio signal, an audio signal containing the schedule creation user intention is cut from the first audio signal to obtain a second audio signal, and then the first schedule information can be generated according to the second audio signal, so that the first schedule information is generated from the audio signal containing the schedule creation user intention cut from the first audio signal, rather than from all the audio signals in the first audio signal, thereby saving the computing resources of the electronic device.

[0009] In some embodiments of the present application, the performing semantic recognition on the first audio signal to obtain a first recognition result includes: performing semantic recognition on the first audio signal to obtain a time intention confidence and an event intention confidence; The time intention confidence is a probability of identifying the first schedule time from the first audio signal; the matter intention confidence is a probability of identifying the first schedule matter from the first audio signal; and the first schedule information includes the first schedule matter and the first schedule time. The method further includes: In a case where a maximum value in the time intention confidence and the matter intention confidence is greater than a first threshold, determining that the first recognition result indicates that a schedule creation user intention is recognized based on the first audio signal.

[0010] In this way, by identifying the time intention confidence of the first schedule time from the first audio signal and the matter intention confidence of the first schedule matter from the first audio signal, whether the first recognition result recognizes the schedule creation user intention based on the first audio signal is determined, that is, in a case where the user does not have any active operation and does not have a wake-up word, the electronic device actively captures the key information such as the schedule time and the schedule matter in the audio signal by perceiving the environmental audio signal in real time, and whether the schedule creation user intention is recognized in the first audio signal is determined by the time intention confidence and the matter intention confidence, thereby avoiding the influence of a single intention on the recognition result of whether the schedule creation user intention is recognized in the first audio signal. If there is only the time intention and no matter intention, the schedule may be only a schedule time and have no specific meaning, and thus the first schedule information is misrecognized.

[0011] In some embodiments of the present application, the second audio signal is obtained by intercepting the audio signal containing the schedule creation user intention from the first audio signal, and includes: The third audio signal is an audio signal including the first schedule time and the first schedule matter. The second information associated with the first information is identified from the first audio signal, and the first information includes at least one of the following: the first schedule time; and the first schedule matter. The fourth audio signal corresponding to the second information is intercepted from the first audio signal. The third audio signal and the fourth audio signal are synthesized to obtain the second audio signal.

[0012] Therefore, by taking the third audio signal containing the first schedule time and the first schedule item in the first audio signal and the fourth audio signal corresponding to the second information associated with at least one of the first schedule time and the first schedule item as the second audio signal for generating the first schedule information, the omission of the schedule information caused by generating the first schedule information only by the third audio signal can be avoided, and the problem that the generated first schedule information is incorrect can be avoided. The scheme of the embodiment of the present application improves the accuracy of the generation of the first schedule information.

[0013] In some embodiments of the present application, the semantic recognition of the first audio signal obtains a time intention confidence and an item intention confidence, including: The first audio signal is input into a first semantic analysis model, and the first audio signal is recognized by the first semantic analysis model to obtain a time intention confidence and an item intention confidence.

[0014] Therefore, whether the first schedule time and the first schedule item are in the first audio signal is recognized by the first semantic analysis model, instead of being recognized manually, which improves the recognition efficiency and objectivity of the first schedule time and the first schedule item in the first audio signal.

[0015] In some embodiments of the present application, after the first audio signal is obtained, the method further includes: The first audio signal is subjected to echo cancellation processing to obtain a fifth audio signal; The fifth audio signal is subjected to sound source positioning according to the positions of the microphones in the electronic device to obtain a sixth audio signal; The sixth audio signal is subjected to noise reduction processing to obtain a seventh audio signal; The semantic recognition of the first audio signal obtains a first recognition result, including: The seventh audio signal is subjected to semantic recognition to obtain a first recognition result.

[0016] Therefore, by combining the microphone array hardware of the electronic device, using noise reduction and beamforming technology, the first audio signal is subjected to noise reduction and enhancement processing before the audio signal enters the model, and the clean audio signal after the noise reduction and enhancement processing can be processed in subsequent processing. Therefore, the influence of noise on the recognition of the audio signal is avoided, and the recognition accuracy of the audio signal in a noisy environment is improved.

[0017] In some embodiments of the present application, the electronic device includes N microphones; N is an integer greater than 1; The sound source positioning of the fifth audio signal according to the positions of the microphones in the electronic device to obtain a sixth audio signal includes: According to a time difference that each two microphones in the electronic device receive the fifth audio signal, M time differences are obtained; According to the M time differences, the fifth audio signal received by the N microphones is weighted and calculated to obtain a sixth audio signal.

[0018] In this way, in the case that the electronic device includes N microphones, M time differences can be obtained according to a time difference that each two microphones in the electronic device receive the fifth audio signal, and then the fifth audio signal received by the N microphones is weighted and calculated according to the M time differences to obtain a sixth audio signal. The sixth audio signal obtained in this way is an audio signal focused on the sound source position, and the audio signal in other directions is suppressed, avoiding the interference of the audio signal in other directions on the subsequent processed audio signal, and improving the accuracy of subsequent identification of the audio signal.

[0019] In some embodiments of the present application, the generating first schedule information according to the second audio signal comprises: performing text recognition on the second audio signal to obtain audio text information corresponding to the second audio signal; obtaining device placement state information, device display screen state information, device location information, and first information; wherein the device placement state information includes a flat state; the device display screen state information includes a lock screen state or a screen-off state; the device location information includes a geographic location of the electronic device; and the first information is used to indicate whether the first schedule is an already created schedule; inputting the audio text information, the device placement state information, the device display screen state information, the device location information, and the first information into a second semantic analysis model, and obtaining a user intention confidence, a first scene confidence, and a second scene confidence through the second semantic analysis model; the user intention confidence is used to indicate a probability of recognizing first text information in the audio text information; the first text information includes the first schedule time and the first schedule item; the first scene confidence is used to indicate a probability that the device placement state information is in a flat state; and the second scene confidence is used to indicate a probability that a first condition is met; wherein the first condition is that the device display screen state information is in a lock screen state or a screen-off state, the device location information is a reference geographic location, and the first information indicates that the first schedule is an already created schedule; performing weighted calculation on the first scene confidence and the second scene confidence through the second semantic analysis model to obtain a scene gating value; calculating the product of the scene gating value and the user intention confidence through the second semantic analysis model to determine a schedule generation decision score;​ The second semantic analysis model is used to generate a schedule generation decision score according to the schedule, and generate a first instruction indicating whether to generate a first schedule; the first instruction indicates to generate the first schedule when the schedule generation decision score is greater than a second threshold; and the first instruction indicates to skip generating the first schedule when the schedule generation decision score is less than or equal to the second threshold. When the first instruction indicates to generate the first schedule, first schedule information is generated according to the first text information.

[0020] In this way, by simultaneously considering whether the user intent confidence of recognizing the first text information containing the first schedule time and the first schedule item in the audio text information, and the device placement state information, the device display screen state information, the device location information, and whether the user's current execution schedule is in the created schedule, etc. The physical scene is used to determine whether to generate the first schedule information, that is, the generation of the first schedule information is determined by the deep combination of the physical scene and the user intent. The accuracy of the first audio signal is improved. The generation accuracy of the first schedule information is improved, and the false positive rate is reduced. In the embodiment of the present application, the second semantic analysis model is used to process the audio text information, the device placement state information, the device display screen state information, the device location information, and whether the user's current execution schedule is in the created schedule, to obtain a schedule generation decision score, and then generate a first instruction indicating whether to generate a first schedule according to the schedule generation decision score. Instead of manually analyzing the audio text information, the device placement state information, the device display screen state information, the device location information, and whether the user's current execution schedule is in the created schedule to determine whether to generate the first schedule, the generation efficiency and objectivity of the first instruction indicating whether to generate the first schedule are improved.

[0021] In some embodiments of the present application, the audio text information includes a first schedule time; the first window further includes a first control; and the first control is used to determine whether to create the first schedule. The method further includes: When the first schedule time is the same as a second schedule time of a second schedule, second schedule information of the second schedule is displayed in the first window. A first input to the first control is received. In response to the first input, the first schedule is created, and the second schedule is deleted.

[0022] In this way, in the case of a conflict between the first schedule and the second schedule, if the user chooses to create the first schedule, the first schedule can be created and the second schedule can be deleted, so that the repeated recording of the schedule at the same schedule time can be avoided.

[0023] In some embodiments of the present application, the displaying the first window comprises: Displaying the first window on a lock screen interface of the electronic device.

[0024] In this way, by displaying the first window on the lock screen interface of the electronic device, and by pushing the first schedule information to the user on the lock screen interface through the silent interaction interface, the user can be directly prompted whether to create the first schedule without unlocking the screen in the case that the display screen of the electronic device is in an off-screen state, without interrupting the operation in the user's hand, and only one click is needed to complete the confirmation, which is zero disturbance, low power consumption, and protects the safety of user privacy data throughout the process.

[0025] In some embodiments of the present application, the first window further comprises a first control and a second control; the second control is used to skip creating the first schedule; After displaying the first window, the method further comprises: In the case that the input of the first control is received, the reward value of the first semantic analysis model is determined as a first reward value; the model parameters of the first semantic analysis model are updated according to the first reward value and a first probability, the first probability being a probability of creating the first schedule based on receiving the first audio signal; In the case that the input of the second control is received, the reward value of the first semantic analysis model is determined as a second reward value; the model parameters of the first semantic analysis model are updated according to the second reward value and a second probability, the second probability being a probability of skipping creating the first schedule based on receiving the first audio signal; Wherein, the first reward value is greater than the second reward value.

[0026] In this way, by responding to the different inputs of the user on the first control and the second control, different reward values for updating the model parameters of the first semantic analysis model can be determined, and then the model parameters of the first semantic analysis model can be updated according to the reward values, so that the model parameters of the first semantic analysis model can be updated differently according to the different choices of the user on the first control and the second control, and the processing result of the first semantic analysis model can be more in line with the personalized use habits of the user.

[0027] In some embodiments of the present application, the obtaining the first audio signal comprises: In the case that the electronic device is in a flat state, the display screen of the electronic device is in an off-screen state, the human voice signal is monitored within a first time length, and a third probability is greater than a third threshold value, the first audio signal is obtained; Wherein, the third probability is used to indicate the probability of monitoring the human voice signal within the first time length.

[0028] Thus, when the electronic device is placed flat, the screen is off, a human voice signal is detected within the first duration, and the third probability of detecting the human voice signal within the first duration is greater than the third threshold, the first audio signal is then acquired. In this way, the first audio signal is acquired only when the conversation is confirmed to be in a high-confidence scenario, and then the subsequent processing flow is executed. In conversations that are not in a high-confidence scenario, it is not necessary to acquire the first audio signal, thus saving the computing resources and power consumption of the electronic device.

[0029] Secondly, embodiments of this application provide an information processing apparatus, which includes: An acquisition module is used to acquire a first audio signal, wherein the first audio signal includes a human voice signal; The display module is used to display a first window when the user's intention to create a schedule is recognized based on the first audio signal; wherein the first window includes first schedule information, first calendar information and first prompt information; The first schedule information includes descriptive information of the schedule elements of the first schedule; the schedule content of the first schedule is determined based on the human voice signal; the first prompt information is used to prompt whether to create the first schedule; the first calendar information includes the first schedule time; the first schedule time is determined based on the human voice signal.

[0030] In some embodiments of this application, the apparatus further includes: The recognition module is used to perform semantic recognition on the first audio signal after acquiring the first audio signal to obtain a first recognition result; the first recognition result is used to indicate whether the user intent to create a schedule is recognized based on the first audio signal. The interception module is used to, when the first recognition result indicates that a user intent to create a schedule is recognized based on the first audio signal, intercept an audio signal containing the user intent to create a schedule from the first audio signal to obtain a second audio signal; The generation module is used to generate first schedule information based on the second audio signal.

[0031] In some embodiments of this application, the identification module is specifically used for: Semantic recognition is performed on the first audio signal to obtain the time intent confidence score and the event intent confidence score; Wherein, the time intent confidence is the probability of identifying the first schedule time from the first audio signal; the event intent confidence is the probability of identifying the first schedule event from the first audio signal; the first schedule information includes the first schedule event and the first schedule time; The device further includes: The determination module is configured to determine, when the maximum value of the time intent confidence score and the event intent confidence score is greater than a first threshold, a first recognition result indication that a user intent to create a schedule was identified based on the first audio signal.

[0032] In some embodiments of this application, the interception module is specifically used for: A third audio signal is extracted from the first audio signal; the third audio signal is an audio signal that includes the first schedule time and the first schedule item. From the first audio signal, identify second information associated with the first information; wherein the first information includes at least one of the following: the first schedule time; the first schedule item; Extract the fourth audio signal corresponding to the second information from the first audio signal; The third audio signal and the fourth audio signal are combined to obtain the second audio signal.

[0033] In some embodiments of this application, the identification module is specifically used for: The first audio signal is input into the first semantic analysis model. The first semantic analysis model is used to perform semantic recognition on the first audio signal to obtain the confidence level of time intent and the confidence level of event intent.

[0034] In some embodiments of this application, the apparatus further includes: The signal removal module is used to perform echo cancellation processing on the first audio signal after acquiring the first audio signal to obtain the fifth audio signal; The sound source localization module is used to locate the sound source of the fifth audio signal based on the position of the microphone in the electronic device, so as to obtain the sixth audio signal; The signal removal module is also used to perform noise reduction processing on the sixth audio signal to obtain the seventh audio signal; The identification module is specifically used for: Semantic recognition is performed on the seventh audio signal to obtain a first recognition result.

[0035] In some embodiments of this application, the electronic device includes N microphones; N is an integer greater than 1; The sound source localization module is specifically used for: Based on the time difference between each pair of microphones receiving the fifth audio signal in the electronic device, M time differences are obtained; ; Based on the M time differences, the fifth audio signal received by the N microphones is weighted and calculated to obtain the sixth audio signal.

[0036] In some embodiments of this application, the generation module is specifically used for: Perform text recognition on the second audio signal to obtain the audio text information corresponding to the second audio signal; The system acquires device placement status information, device display status information, device location information, and first information; wherein, the device placement status information includes a flat position; the device display status information includes a locked screen or a screen-off state; the device location information includes the geographical location of the electronic device; and the first information is used to indicate whether the user's currently executing schedule is a previously created schedule. The audio text information, device placement status information, device display status information, device location information, and first information are input into a second semantic analysis model. Through the second semantic analysis model, user intent confidence, first scenario confidence, and second scenario confidence are obtained. The user intent confidence indicates the probability of recognizing the first text information in the audio text information. The first text information includes the first schedule time and the first schedule item. The first scenario confidence indicates the probability that the device placement status is in a flat position. The second scenario confidence indicates the probability that a first condition is met. The first condition is that the device display status is locked or off, the device location information is a reference geographical location, and the first information indicates that the user's currently executing schedule is a created schedule. The first scenario confidence score and the second scenario confidence score are weighted and calculated using the second semantic analysis model to obtain the scenario gating value. The product of the context gating value and the user intent confidence level is calculated using the second semantic analysis model to determine the schedule generation decision score. Using the second semantic analysis model, a first instruction is generated based on the schedule generation decision score; the first instruction is used to indicate whether to generate the first schedule; if the schedule generation decision score is greater than a second threshold, the first instruction indicates to generate the first schedule; if the schedule generation decision score is less than or equal to the second threshold, the first instruction indicates to skip generating the first schedule. When the first instruction instructs the generation of a first schedule, first schedule information is generated based on the first text information.

[0037] In some embodiments of this application, the audio text information includes a first schedule time; the first window further includes a first control; the first control is used to determine the creation of the first schedule; The display module is further configured to: display the second schedule information of the second schedule in the first window when the first schedule time is the same as the second schedule time of the second schedule; The device further includes: A receiving module is used to receive the first input to the first control; A schedule processing module is used to create the first schedule and delete the second schedule in response to the first input.

[0038] In some embodiments of this application, the display module is specifically used for: The first window is displayed on the lock screen of the electronic device.

[0039] In some embodiments of this application, the first window further includes a first control and a second control; the second control is used to skip the creation of the first schedule; The device further includes: The model parameter update module is configured to, after displaying the first window and upon receiving input to the first control, determine that the reward value of the first semantic analysis model is a first reward value; update the model parameters of the first semantic analysis model according to the first reward value and a first probability, wherein the first probability is the probability of creating the first schedule based on receiving the first audio signal; determine that the reward value of the first semantic analysis model is a second reward value upon receiving input to the second control; update the model parameters of the first semantic analysis model according to the second reward value and a second probability, wherein the second probability is the probability of skipping the creation of the first schedule based on receiving the first audio signal; wherein the first reward value is greater than the second reward value.

[0040] In some embodiments of this application, the acquisition module is specifically used for: The first audio signal is acquired when the electronic device is placed flat, the screen of the electronic device is off, a human voice signal is detected within the first time period, and the third probability is greater than the third threshold. The third probability is used to indicate the probability of detecting a human voice signal within the first time period.

[0041] Thirdly, embodiments of this application provide an electronic device, which includes a processor and a memory, wherein the memory stores programs or instructions that can run on the processor, and the programs or instructions, when executed by the processor, implement the method as described in the first aspect.

[0042] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the method described in the first aspect.

[0043] Fifthly, embodiments of this application provide a chip, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.

[0044] In a sixth aspect, embodiments of this application provide a computer program product stored in a storage medium, which is executed by at least one processor to implement the method described in the first aspect.

[0045] In this embodiment, when the user's intention to create a schedule is identified in the first audio signal containing human voice signals, a first window can be displayed. This first window includes first schedule information, first calendar information, and a first prompt. The first schedule information includes descriptive information of the schedule elements, and the schedule content is determined based on the human voice signal. The first prompt is used to prompt whether to create the first schedule. The first calendar information includes the first schedule time, which is also determined based on the human voice signal. Thus, the electronic device can automatically and accurately extract core elements such as time and events from continuous human voice signals and automatically fill them into the first window. This eliminates the need for the user to listen to the entire recording on the electronic device or stop their current task to create the schedule. It also saves the user from a series of tedious operations such as manually opening the calendar application, entering text one by one, and selecting dates. The user only needs to confirm the first schedule information generated in the first window, simplifying multiple steps into a single confirmation. This greatly shortens the schedule creation path and time, improves the accuracy and efficiency of schedule creation, and avoids interruptions to the user's current task in serious, immersive scenarios such as business meetings due to the need to manually create a schedule.

[0046] Understandably, the information processing apparatus provided in the second aspect, the electronic device provided in the third aspect, the readable storage medium provided in the fourth aspect, the chip provided in the fifth aspect, and the computer program product provided in the sixth aspect are all used to execute the method provided in this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating the information processing method provided in some embodiments of this application; Figure 2 This is a schematic diagram showing the first window provided in some embodiments of this application; Figure 3 This is a schematic diagram showing the first window provided in some embodiments of this application; Figure 4 This is a flowchart illustrating the information processing method provided in some embodiments of this application; Figure 5 These are schematic diagrams illustrating the structure of an information processing apparatus according to some embodiments of this application; Figure 6 These are schematic diagrams illustrating the structure of an electronic device according to some embodiments of this application; Figure 7 These are schematic diagrams illustrating the hardware structure of an electronic device according to some embodiments of this application. Detailed Implementation

[0048] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0049] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and are not limited in number; for example, a first object can be one or N objects. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0050] The terminology used in the embodiments of this invention will be explained below.

[0051] Large Language Models (LMMs) are deep learning models trained on massive amounts of text data, enabling them to generate natural language text or understand the meaning of language text. These models can provide in-depth knowledge and language production on a wide range of topics through training on large datasets. Their core idea is to learn patterns and structures of natural language through large-scale unsupervised training, mimicking human language cognition and generation processes to some extent.

[0052] Automatic Speech Recognition (ASR) is a system that converts human speech into text or commands through techniques such as feature extraction, pattern matching, and model training. Its workflow consists of two stages: acoustic model construction and language model application. The acoustic model analyzes speech signals to extract phonemes, frequencies, and other features, and then trains the recognition system. The language model uses statistical methods to predict the probability and semantics of word sequences. The combination of these two approaches achieves the speech-to-text conversion.

[0053] Neural Processing Unit (NPU): A dedicated acceleration chip designed specifically for neural network computing. Its core advantage lies in its optimized hardware architecture, which efficiently processes core neural network operations such as matrix operations and convolution calculations. Compared with general-purpose central processing units, NPUs have higher energy efficiency and lower latency, making them particularly suitable for real-time artificial intelligence (AI) tasks.

[0054] Digital Signal Processor (DSP): A programmable chip specifically designed for digital signal processing.

[0055] An Inertial Measurement Unit (IMU) is a device that measures an object's three-axis attitude angles, angular rates, and accelerations. An IMU contains three single-axis accelerometers and three single-axis gyroscopes. The accelerometers detect the object's acceleration signals along the three independent axes of the carrier's coordinate system, while the gyroscopes detect the carrier's angular velocity signals relative to the navigation coordinate system. By measuring the object's angular velocity and acceleration in three-dimensional space, the object's attitude can be calculated.

[0056] Voice activity detection (VAD) is a technique used in speech processing to detect the presence of a speech signal. In this embodiment, it is used to detect the presence of a human voice.

[0057] Schedule: refers to a plan or event that is strongly tied to a specific point in time or time period, the core of which is "when and what will happen".

[0058] User intent for schedule creation: This refers to a user expressing to the system, through explicit or implicit instructions, their desire to complete a planned activity or to-do item at a specific point in the future or within a specific time period, and recording that planned activity or to-do item to be completed at that specific point in the future or within a specific time period.

[0059] Schedule elements: These refer to all the basic information units or attributes that constitute a complete schedule event. They are used to clearly and unambiguously describe the specific schedule items, schedule times, schedule locations, and related contexts of a schedule. Here, schedule time refers to a specific point in time or time period within the schedule, and schedule items refer to activities or tasks planned to be completed within the schedule.

[0060] Semantic analysis models are a type of artificial intelligence model whose core task is to enable computers to understand the "meaning" expressed by human language. They aim to go beyond the analysis of superficial structures such as vocabulary and grammar, thereby gaining insight into the intentions, concepts, emotions, and contextual relationships behind language.

[0061] Confidence level: A quantitative metric that measures the degree of certainty or probability that a judgment, prediction, or proposition is true. It is usually expressed as a percentage or a decimal between 0 and 1. In AI, confidence level typically refers to the degree of confidence a model has in the correctness of its predictions.

[0062] The technical solution of this application embodiment can be applied to scenarios where to capture and record schedule and to-do information in fleeting moments. For example, at 10:00 AM on November 29, 2025, a user is to accompany Sales Manager B from Company A to visit Company C and conduct business negotiations with Project Manager D from Company C. After the user and Sales Manager B arrive at the conference room of Company C, the user places their mobile phone flat on the table and locks the screen. During the meeting, Sales Manager B and Project Manager D have a detailed discussion on a certain project. Sales Manager B discusses the details of the project plan, and Project Manager D comments that the plan is very good. After the project plan is discussed, both parties finalize the intention to cooperate. At the end of the meeting, Project Manager D says to Sales Manager B, "Come to my company for a meeting next Tuesday afternoon. Please bring the contract." Sales Manager B then asks, "What time in the afternoon?" Project Manager D replies, "3:00 PM." Sales Manager B asks, "Do you need to bring any other materials?" Project Manager D replies, "Also, bring the samples." For example, at 2 PM on November 30, 2025, a user is attending a product communication session at an electronics product launch event. The schedule for this product communication session is pre-recorded in the user's phone's calendar app. Additionally, the user's calendar app also shows a scheduled group meeting at 3 PM on December 2, 2025. After entering the product communication session venue, the user finds their seat, places their phone flat on the seat, and locks the screen. During the product communication session, the host introduces the new electronic device and its release date, for example, saying, "This phone will be available for purchase at 0:00 on December 31, 2025."

[0063] The information processing method provided in this application will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.

[0064] Figure 1This is a flowchart illustrating an information processing method provided in an embodiment of this application. The subject executing this information processing method can be an electronic device, which may be, but is not limited to, a personal computer (PC), a smartphone, a tablet computer, or a personal digital assistant (PDA).

[0065] like Figure 1 As shown, the information processing method provided in this application embodiment may include steps 110-120.

[0066] Step 110: Obtain the first audio signal.

[0067] The first audio signal can be an audio signal acquired during user conversations, meaning it includes human voice signals. For example, the first audio signal could be the audio signal of all conversations between Sales Manager B and Project Manager D in the meeting mentioned above. As another example, the first audio signal could also be all audio signals from a product communication meeting attended by the user, as mentioned above.

[0068] In some embodiments of this application, the first audio signal can be acquired in real time by an electronic device, that is, during a meeting between sales manager B and project manager D, the content of their conversation can be acquired in real time. Another example is acquiring the content of the presenter's introduction to the released electronic device in real time during a product communication meeting.

[0069] In some embodiments of this application, the first audio signal is acquired only when the electronic device is determined to be in a preset session scenario. When the electronic device is not in a preset session scenario, it is not necessary to acquire the first audio signal, thus saving the computing resources of the electronic device.

[0070] The aforementioned preset conversation scenarios can be conversation scenarios that require analysis of conversation information, such as conference scenarios, exhibition scenarios, etc.

[0071] In some embodiments of this application, in order to save the computing resources and power consumption of the electronic device, step 110 may specifically include: The first audio signal is acquired when the electronic device is placed flat, the screen of the electronic device is off, a human voice signal is detected within the first time period, and the third probability is greater than the third threshold.

[0072] In this context, "laying flat" can refer to an electronic device being placed horizontally. For example, in the example above, after Sales Manager B arrives at Company C's meeting room, he places his phone flat on the table and locks it; in this case, the user's phone is in a laying flat state. As another example, in the example above, after entering the product communication meeting room, the user finds their seat, places their phone flat on the seat, and locks it; in this case, the user's phone is in a laying flat state.

[0073] The first duration can be a pre-set duration for monitoring human voice signals. For example, the value of the first duration can be 3-10 seconds, such as 4 seconds. The specific value of the first duration can be set by the user according to their needs, and is not limited in this embodiment.

[0074] The third probability can be used to indicate the probability of detecting a human voice signal within the first time period. The value of the third probability can be between 0 and 1, for example, the third probability can be 0.2, 0.5 or 0.8, etc.

[0075] The third threshold can be a pre-set threshold for a third probability, and the value of the third threshold ranges from 0.6 to 1. For example, the third threshold can be 0.75, 0.8, or 0.9, etc. The specific value of the third threshold can be set by the user according to their needs, and is not limited in this embodiment.

[0076] In some embodiments of this application, when the electronic device is detected to be in a flat position, the screen of the electronic device is in a screen-off state, a human voice signal is detected within a first duration, and the third probability of detecting the human voice signal within the first duration is greater than a third threshold, it indicates that the electronic device is in a preset conversation scenario, and the first audio signal can be acquired at this time.

[0077] The determination of whether an electronic device is in a flat position can be obtained through the IMU module in the electronic device. That is, the IMU detects the pose information of the electronic device, and the pose information can be used to determine whether the electronic device is in a flat position. If the IMU detects that the electronic device is in a flat position, the placement status information of the electronic device can be marked as the value "1".

[0078] The determination of whether the screen of an electronic device is in a screen-off state can be obtained through the system detection module of the electronic device. Specifically, the system detection module can detect the brightness of the screen of the electronic device. If the brightness of the screen of the electronic device is detected to be lower than a certain threshold, such as the brightness of the screen of the electronic device being lower than 1 nit, then it is determined that the screen of the electronic device is in a screen-off state, and the screen status information of the electronic device can be marked with the value "1".

[0079] Sound detection in electronic devices can be achieved through the VAD (Voice Awareness) module. For example, if the VAD module detects a human voice signal lasting more than 4 seconds with a probability higher than 0.75, then the electronic device is confirmed to have detected a human voice signal, and the VAD status of the electronic device can be marked as "1".

[0080] If the electronic device's placement status information is identified as "1", the electronic device's screen status information is identified as "1", and the electronic device's VAD status is identified as "1", then the electronic device can be determined to be in a preset conversation scenario. The electronic device will then trigger the schedule to-do record function, at which point the first audio signal can be obtained.

[0081] Continuing with the first example above, after Sales Manager B arrives at Company C's meeting room, he places his mobile phone flat on the table and locks the screen. The IMU detects that the phone is flat, and the system detection module detects that the screen is off. Then, Sales Manager B and Project Manager D begin a detailed meeting about a project. If the VAD module detects a voice signal lasting more than 5 seconds, and the probability of detecting such a signal exceeds 0.75, the microphone of the electronic device can be activated to continuously capture the conversation between Sales Manager B and Project Manager D during the meeting.

[0082] In some embodiments of this application, when the electronic device is in a flat position, the display screen of the electronic device is in a screen-off state, a human voice signal is detected within a first duration, and the third probability of detecting the human voice signal within the first duration is greater than a third threshold, the first audio signal is then acquired. In this way, the first audio signal is acquired only when it is confirmed to be in a high-confidence conversation scenario, and then the subsequent processing flow is executed. In conversation scenarios that are not in a high-confidence scenario, it is not necessary to acquire the first audio signal, thus saving the computing resources and power consumption of the electronic device.

[0083] Step 120: If the user intent to create a schedule is recognized based on the first audio signal, the first window is displayed.

[0084] In this context, the user intent for creating a schedule can be the user's intention to create a schedule. For example, in the first example above, after project manager D tells the user to bring the contract and samples to the other party's company next Tuesday at 3 PM, the user wants to create a schedule for a meeting at project manager D's company next Tuesday at 3 PM. Therefore, the user's intention to create a schedule for a meeting at project manager D's company next Tuesday at 3 PM is a schedule creation user intent. As another example in the second example above, after the host announces that the phone will be available for purchase at 0:00 on December 31, 2025, the user wants to create a schedule to remind themselves to buy the phone then. Therefore, the user's intention to create a schedule for the new phone to be available for purchase at 0:00 on December 31, 2025 is a schedule creation user intent.

[0085] The first window can be a window displayed on the screen of an electronic device. This first window may include first schedule information, first calendar information, and first notification information. The aforementioned first schedule information may include descriptive information about the schedule elements of the first schedule. This first schedule is the schedule the user wants to create, and its content is determined based on human voice signals. Schedule elements may include the schedule time and schedule events.

[0086] The aforementioned first prompt information is used to prompt whether to create a first schedule; the first calendar information includes the first schedule time, which can be the schedule time of the first schedule, and the first schedule time is determined based on human voice signals.

[0087] Continuing with the first example above, if the user's desired schedule is determined based on the first audio signal, indicating they want to meet with project manager D at 3 PM next Tuesday with a contract and samples, then... Figure 2 As shown, a first window 21 is displayed, which includes first schedule information 211, first calendar information 212 and first prompt information 213.

[0088] The first schedule information 211 includes a description of the schedule elements for creating a meeting at Project Manager D's company next Tuesday at 3 PM, where participants will bring the contract and samples. These schedule elements are: the first schedule time: next Tuesday at 3 PM; and the first schedule item: bringing the contract and samples. This can be seen in the following example. Figure 2 The first agenda item 214, "Go to Project Manager D's company for a meeting next Tuesday at 3 PM with the contract and samples," was determined based on a conversation between Project Manager D and Sales Manager B during the meeting.

[0089] Figure 2 The first prompt message 213, "Do you want to replace the schedule with a new one-click message?", is used to prompt whether to create the first schedule 214. Figure 2The first calendar information 212 includes the first schedule time "2025.12.2", which is also determined based on the conversation between Project Manager D and Sales Manager B in the meeting.

[0090] Continuing with the second example above, if the user is identified as wanting to create a schedule for the new phone to go on sale at 00:00 on December 31, 2025, based on the first audio signal, it can be done as follows: Figure 3 As shown, a first window 31 is displayed, which includes first schedule information 311, first calendar information 312 and first prompt information 313.

[0091] The first schedule information 311 includes a description of the schedule elements for the new mobile phone to be launched at 00:00 on December 31, 2025. These schedule elements are the first schedule time: 00:00 on December 31, 2025, and the first schedule item: launch of the new mobile phone. This can be seen in the following... Figure 3 The first schedule, 314, states that "the new mobile phone will go on sale at 0:00 on December 31, 2025." The content of the first schedule, 314, was determined based on what the host said during the product communication meeting.

[0092] Figure 3 The first prompt message 313, "Do you want to record a new schedule with one click?", is used to prompt whether to create the first schedule 314. Figure 3 The first calendar information 312 includes the first schedule time "2025.12.31", which was determined based on what the host said during the product communication meeting.

[0093] In some embodiments of this application, in order to save computing resources of electronic devices, after step 110, the method described above may further include: Perform semantic recognition on the first audio signal to obtain the first recognition result; If the first recognition result indicates that the user intent to create a schedule is recognized based on the first audio signal, the second audio signal is obtained by extracting the audio signal containing the user intent to create a schedule from the first audio signal. The first schedule information is generated based on the second audio signal.

[0094] The first recognition result is used to indicate whether the user intent to create the schedule is recognized based on the first audio signal.

[0095] The second audio signal can be an audio signal extracted from the first audio signal that identifies the user's intention to create the schedule. For example, in the first example above, when project manager D tells sales manager B, "Come to my company for a meeting next Tuesday afternoon, and bring the contract," the user wants to create a schedule for a meeting at project manager D's company next Tuesday afternoon, bringing the contract. In this case, the second audio signal is the audio signal spoken by project manager D, "Come to my company for a meeting next Tuesday afternoon, and bring the contract." As another example, in the second example above, when the host says, "This phone will be available for sale at 0:00 on December 31, 2025," the user wants to create a schedule for the new phone to be available for sale at 0:00 on December 31, 2025. In this case, the second audio signal is the audio signal spoken by the host, "This phone will be available for sale at 0:00 on December 31, 2025."

[0096] In some embodiments of this application, semantic recognition can be performed on the first audio signal first. If the user intent to create a schedule is recognized based on the first audio signal, an audio signal including the user intent to create a schedule can be extracted from the first audio signal to obtain a second audio signal. Then, the first schedule information can be generated based on the second audio signal.

[0097] Continuing with the first example above, semantic recognition is performed on the conversation between Project Manager D and Sales Manager B. When Project Manager D says to Sales Manager B, "Come to my company for a meeting next Tuesday afternoon, and bring the contract," the user intent to create a schedule is identified. At this point, the audio signal "Come to my company for a meeting next Tuesday afternoon, and bring the contract" can be extracted to obtain the second audio signal. Based on this second audio signal, a schedule creation function can be generated. Figure 2 The first schedule information 211 is shown.

[0098] Continuing with the second example above, semantic recognition is performed on the host's entire conversation. When the host says, "This phone will be available for purchase at 00:00 on December 31, 2025," the user intent to create a schedule is identified. At this point, the audio signal "This phone will be available for purchase at 00:00 on December 31, 2025" can be extracted to obtain the second audio signal. Based on this second audio signal, a schedule creation function can be generated. Figure 3 The first schedule information 311 is shown.

[0099] In the embodiments of this application, when the user intent to create a schedule is identified based on the first audio signal, an audio signal including the user intent to create a schedule is extracted from the first audio signal to obtain a second audio signal. Then, the first schedule information can be generated based on the second audio signal. In this way, the first schedule information is generated from the audio signal including the user intent to create a schedule extracted from the first audio signal, rather than generating the first schedule information based on all the audio signals in the first audio signal, thus saving the computing resources of the electronic device.

[0100] In some embodiments of this application, the step of performing semantic recognition on the first audio signal to obtain a first recognition result may specifically include: Semantic recognition is performed on the first audio signal to obtain the confidence scores for time intent and event intent. The methods mentioned above may also include: If the maximum value of the time intent confidence score and the event intent confidence score is greater than a first threshold, the first recognition result indicates that the user intent to create the schedule was identified based on the first audio signal.

[0101] The first schedule information can include the first schedule item and the first schedule time. In the first example above, the first schedule item is going to Project Manager D's company with the contract, and the first schedule time is next Tuesday afternoon. In the second example above, the first schedule item is the launch of a new mobile phone, and the first schedule time is 00:00 on December 31, 2025.

[0102] The time intent confidence level is the probability of identifying the first scheduled time from the first audio signal. For example, in the first example above, the time intent confidence level is the probability of identifying next Tuesday afternoon from the conversation between Sales Manager B and Project Manager D. Similarly, in the second example above, the time intent confidence level is the probability of identifying 00:00 on December 31, 2025 from the host's words.

[0103] The event intent confidence level is the probability of identifying the first scheduled event from the first audio signal. For example, in the first example above, the event intent confidence level is the probability of identifying a contract from the conversation between Sales Manager B and Project Manager D. Similarly, in the second example above, the event intent confidence level is the probability of identifying the launch of a new mobile phone from the host's words.

[0104] The first threshold can be the maximum value of the pre-set confidence level of the event intent and the confidence level of the time intent. For example, the value range of the first threshold can be 0.6 to 1, such as 0.65, 0.8 or 0.95. The specific value of the first threshold can be set by the user according to their needs, and is not limited in this embodiment.

[0105] In some embodiments of this application, semantic recognition can be performed on the first audio signal to obtain the confidence level of the time intent of the first schedule time identified from the first audio signal, and the confidence level of the event intent of the first schedule item identified from the first audio signal. Then, if the maximum value of the time intent confidence level and the event intent confidence level is greater than a first threshold, it can be determined that the first recognition result indicates the user intent of schedule creation identified based on the first audio signal.

[0106] Continuing with the example above, taking a first threshold of 0.65 as an example, if the first scheduled time "next Tuesday afternoon" and the first scheduled item "Come to my company for a meeting, please bring the contract" are identified during the entire meeting between Sales Manager B and Project Manager C, and the probability of identifying the first scheduled time "next Tuesday afternoon" is 0.65, and the probability of identifying the first scheduled item "Come to my company for a meeting, please bring the contract" is 0.7, then if the maximum value of 0.65 and probability 0.7 is greater than the first threshold of 0.65, then it can be determined that the user intent to create the schedule was identified during the entire meeting between Sales Manager B and Project Manager C.

[0107] It should be noted that, as described above, when determining that the first identification result indicates that a user intent to create a schedule has been identified based on the first audio signal, this is only determined when the maximum value of the confidence scores for time intent and event intent is greater than a first threshold. However, in other embodiments, the first identification result indicating that a user intent to create a schedule has been identified based on the first audio signal can also be determined based on other statistical information regarding the confidence scores for time intent and event intent. For example, the first identification result indicating that a user intent to create a schedule has been identified based on the first audio signal can be determined when the average value of the confidence scores for time intent and event intent is greater than a first threshold, or when the minimum value of the confidence scores for time intent and event intent is greater than a first threshold. This application does not impose any limitations on this aspect.

[0108] In the embodiments of this application, the first identification result is determined by identifying the time intent confidence of the first schedule time and the event intent confidence of the first schedule item from the first audio signal. That is, without any active operation or wake word from the user, the electronic device uses real-time perception of the ambient audio signal to actively capture key information such as schedule time and schedule item in the audio signal. In this way, the time intent confidence and event intent confidence are used together to determine whether the schedule creation user intent is identified in the first audio signal, avoiding the influence of a single intent on the identification result of whether the schedule creation user intent is identified in the first audio signal. This is because if there is only a time intent and no event intent, the schedule may just be a schedule time without any specific meaning, which would cause misidentification of the first schedule information.

[0109] In some embodiments of this application, to improve the accuracy of the first schedule information generation, the step of extracting an audio signal containing the user's intention to create the schedule from the first audio signal to obtain a second audio signal may specifically include: Extract the third audio signal from the first audio signal; Identify second information associated with the first information from the first audio signal; Extract the fourth audio signal corresponding to the second information from the first audio signal; The third and fourth audio signals are combined to obtain the second audio signal.

[0110] The third audio signal is an audio signal that includes the first schedule time and the first schedule item. In the first example above, the third audio signal is the project manager D saying to the sales manager B, "Come to my company for a meeting next Tuesday afternoon, and bring the contract with you."

[0111] The first piece of information may include at least one of the first scheduled time and the first scheduled item. The second piece of information may be information associated with the first piece of information; that is, the second piece of information is information associated with at least one of the first scheduled time and the first scheduled item. For example, in the first example above, after Project Manager D says, "Come to my company for a meeting next Tuesday afternoon, please bring the contract," Sales Manager B asks, "What time in the afternoon?" Project Manager D replies, "Three o'clock in the afternoon." Therefore, Project Manager D's "Three o'clock in the afternoon" is the second piece of information associated with the first scheduled time, "next Tuesday afternoon," in the statement "Come to my company for a meeting next Tuesday afternoon, please bring the contract." After Project Manager D says, "Come to my company for a meeting next Tuesday afternoon, please bring the contract," Sales Manager B also asks, "Do I need to bring any other materials?" Project Manager D replies, "Also, bring the samples." Therefore, Project Manager D's "Also, bring the samples" is the second piece of information associated with the first scheduled item, "Come to my company for a meeting next Tuesday afternoon, please bring the contract," in the statement "Come to my company for a meeting next Tuesday afternoon, please bring the contract."

[0112] The fourth audio signal can be an audio signal extracted from the first audio signal that corresponds to the second information. Taking the first example above, the audio signals "What time is it this afternoon?" and "3 pm" before "Come to my company for a meeting next Tuesday afternoon, please bring the contract" can be extracted from the dialogue between project manager D and sales manager B, as well as the audio signals "Do you need to bring any other materials?" and "Also, bring some samples" after "Come to my company for a meeting next Tuesday afternoon, please bring the contract". The above-mentioned "What time is it this afternoon?", "3 pm", "Do you need to bring any other materials?", and "Also, bring some samples" are the fourth audio signals.

[0113] In some embodiments of this application, a third audio signal containing the first schedule time and the first schedule item can be extracted from the first audio signal. Then, second information associated with at least one of the first schedule time and the first schedule item can be identified from the first audio signal. Subsequently, a fourth audio signal corresponding to the second information can be extracted from the first audio signal, and the third audio signal and the fourth audio signal can be synthesized to obtain the second audio signal.

[0114] Continuing with the first example above, the third audio signal is "Come to my company for a meeting next Tuesday afternoon, please bring the contract." The fourth audio signal consists of the audio signals before "Come to my company for a meeting next Tuesday afternoon, please bring the contract" ("What time in the afternoon?" and "3 PM"), and the audio signals after "Come to my company for a meeting next Tuesday afternoon, please bring the contract" ("Do you need to bring any other materials?" and "Please also bring samples"). The third and fourth audio signals can be combined to obtain the second audio signal: "Come to my company for a meeting next Tuesday afternoon, please bring the contract. What time in the afternoon? 3 PM. Do you need to bring any other materials? Please also bring samples."

[0115] In the embodiments of this application, by using a third audio signal containing a first schedule time and a first schedule item in the first audio signal, and a fourth audio signal corresponding to second information associated with at least one of the first schedule time and the first schedule item as the second audio signal for generating the first schedule information, the problem of omission of schedule information caused by relying solely on the third audio signal to generate the first schedule information, which in turn leads to errors in the generated first schedule information, can be avoided. The solution of the embodiments of this application improves the accuracy of the generation of the first schedule information.

[0116] In some embodiments of this application, in order to improve the efficiency and objectivity of identifying the first schedule time and the first schedule item in the first audio signal, the step of performing semantic recognition on the first audio signal to obtain the time intent confidence score and the item intent confidence score may specifically include: The first audio signal is input into the first semantic analysis model. The first semantic analysis model performs semantic recognition on the first audio signal to obtain the confidence level of time intent and the confidence level of event intent.

[0117] The first semantic analysis model can be a pre-trained model for performing contextual semantic analysis on the first audio signal. The first semantic analysis model may include, but is not limited to, a neural network model based on deep learning, a support vector machine model, or a decision tree model. The specific type of the first semantic model can be selected according to the user's needs and is not limited in this embodiment.

[0118] It should be noted that the first semantic analysis model can be a lightweight streaming model deployed on the energy efficiency core of the NPU, and the first semantic analysis model can operate at ultra-low power consumption at the milliwatt level.

[0119] In some embodiments of this application, the first audio signal can be input into the first semantic analysis model, and the first semantic analysis model can be used to perform semantic recognition on the first audio signal to obtain the confidence level of time intent and the confidence level of event intent.

[0120] Continuing with the first example above, the first audio signal from the real-time conversation between Sales Manager B and Project Manager D can be input into the first semantic analysis model. Based on the first semantic analysis model, it can be identified whether there is a first schedule time and a first schedule item in the first audio signal from the conversation between Sales Manager B and Project Manager D, thereby obtaining the confidence level of the time intent and the confidence level of the item intent.

[0121] In the embodiments of this application, a first semantic analysis model is used to identify whether there is a first schedule time and a first schedule item in the first audio signal, rather than manually identifying whether there is a first schedule time and a first schedule item in the first audio signal, thereby improving the efficiency and objectivity of the identification of the first schedule time and the first schedule item in the first audio signal.

[0122] In some embodiments of this application, in order to improve the accuracy of recognizing the first audio signal, after step 110, the method described above may further include: Echo cancellation is performed on the first audio signal to obtain the fifth audio signal; Based on the position of the microphone in the electronic device, the sound source of the fifth audio signal is located to obtain the sixth audio signal; The sixth audio signal is denoised to obtain the seventh audio signal; The step of performing semantic recognition on the first audio signal to obtain a first recognition result includes: Semantic recognition was performed on the seventh audio signal to obtain the first recognition result.

[0123] The fifth audio signal can be the audio signal obtained after echo cancellation processing of the first audio signal; that is, the fifth audio signal is the audio signal obtained after removing the echo audio signal from the first audio signal. The aforementioned echo audio signal can be an audio signal emitted by the speaker of an electronic device, reflected by the surrounding environment, and then received by the microphone of the electronic device. For example, in the first example above, during the meeting, Sales Manager B's mobile phone received a news push notification. When the electronic device received the news push notification, its speaker emitted a "beep" sound. This "beep" sound was reflected by the walls of the meeting room, creating an echo, which was then received by the microphone of the electronic device. Therefore, the echo of the "beep" sound is the echo audio signal.

[0124] The sixth audio signal can be the audio signal obtained by locating the sound source of the fifth audio signal based on the position of the microphone in the electronic device.

[0125] The seventh audio signal can be the audio signal obtained after denoising the sixth audio signal; that is, the seventh audio signal can be the audio signal obtained after removing the environmental noise audio signal from the sixth audio signal. The aforementioned environmental noise audio signal can be an audio signal of environmental noise. For example, in the first example above, the air conditioner in the conference room was on throughout the meeting, and the air conditioner would generate noise. The noise generated by the air conditioner is the environmental noise audio signal.

[0126] In some embodiments of this application, for an echoing audio signal, the first audio signal can be first subjected to echo cancellation processing to obtain a fifth audio signal. Then, based on the position of the microphone in the electronic device, the fifth audio signal is located by sound source localization, and the fifth audio signal is enhanced to suppress user noise to obtain a sixth audio signal. Then, the sixth audio signal is subjected to noise reduction processing to obtain a seventh audio signal. In this way, semantic recognition can be performed on the seventh audio signal in subsequent recognition to obtain a first recognition result.

[0127] Continuing with the first example above, in the meeting, Project Manager D is positioned opposite Sales Manager B. While Project Manager D is talking to Sales Manager B, Sales Manager B is also using a laptop to record supplementary information about the solutions mentioned by Project Manager D during the meeting. Thus, after obtaining the audio signal of Project Manager D's conversation with Sales Manager B, we can first eliminate the echo of the "beep" notification sound caused by Sales Manager B's phone receiving a news push notification. Then, we can use the multi-microphone array of the electronic device to calculate the position of Project Manager D, i.e., the sound source location. Then, using this sound source location, we can focus the sound pickup on Project Manager D opposite Sales Manager B, suppressing the sound of the user typing on the keyboard behind the phone, to obtain the sixth audio signal. Finally, we can remove the air conditioning noise from the sixth audio signal to obtain the seventh audio signal.

[0128] In the embodiments of this application, by combining electronic device microphone array hardware and utilizing noise reduction and beamforming technology, the first audio signal is denoised and enhanced before entering the model. In subsequent processing, the clean audio signal after denoising and enhancement can be used for processing, thus avoiding the influence of noise on the recognition of audio signals and improving the accuracy of audio signal recognition in noisy environments.

[0129] In some embodiments of this application, the electronic device may include N microphones, where N is a positive integer greater than 1, i.e., the electronic device is a multi-microphone array electronic device. Thus, the step of locating the sound source of the fifth audio signal based on the positions of the microphones in the electronic device to obtain the sixth audio signal may specifically include: Based on the time difference between each pair of microphones receiving the fifth audio signal in the electronic device, M time differences are obtained; The sixth audio signal is obtained by weighting the fifth audio signal received by N microphones based on M time differences.

[0130] Here, M time differences can be the time differences between every two microphones in the electronic device receiving the fifth audio signal, and the relationship between M and N is as follows: .

[0131] In some embodiments of this application, M time differences can be obtained based on the time difference between each pair of microphones receiving the fifth audio signal in the electronic device, thus determining the direction of the sound source. Then, based on the calculated sound source direction, the fifth audio signals received by all microphones are aligned according to their corresponding time differences. The aim is to ensure that the sound signal from the sound source direction is in phase across all microphone channels. The fifth audio signals received by each microphone channel after the time alignment operation are then superimposed; that is, the fifth audio signals received by N microphones are weighted. Thus, for the audio signal from the sound source direction, since the audio signals from all directions have been aligned to the sound source direction, their phases are the same. Therefore, during superposition, the sound wave peaks and valleys are aligned, and the signal amplitudes are superimposed in phase, significantly enhancing the audio signal. For sounds from other directions, since the noise sources are in different directions, the compensated time delay is not applicable to them. Therefore, their phases are random during superposition, causing them to cancel each other out, and their amplitudes are significantly suppressed. This enhances the pickup of sound from the sound source direction, suppresses noise from other directions, and focuses the pickup on the sound source direction.

[0132] In the embodiments of this application, when the electronic device includes N microphones, M time differences can be obtained based on the time difference between each pair of microphones receiving the fifth audio signal. Then, based on the M time differences, the fifth audio signals received by the N microphones are weighted and calculated to obtain the sixth audio signal. The sixth audio signal obtained in this way is an audio signal focused on the sound source location, suppressing audio signals from other directions, avoiding interference from audio signals from other directions on the audio signals processed subsequently, and improving the accuracy of subsequent audio signal recognition.

[0133] In some embodiments of this application, before performing echo cancellation processing on the first audio signal to obtain the fifth audio signal, the method described above may further include: Echo path prediction is performed on the audio signal emitted by the speaker of an electronic device to obtain the predicted echo audio signal; The echo audio signal is obtained by finding the audio signal that corresponds to the predicted echo audio signal in the first audio signal; The process of performing echo cancellation on the first audio signal to obtain the fifth audio signal includes: The echo audio signal in the first audio signal is removed to obtain the fifth audio signal.

[0134] The predicted echo audio signal can be obtained by performing echo path prediction on the audio signal emitted by the speaker of an electronic device based on the echo path prediction principle. For example, in the above example, during the meeting, Sales Manager B's mobile phone received a news push notification. When the electronic device received the news push notification, the speaker of the electronic device emitted a "beep" notification sound. The echo path prediction principle can be used to predict the echo of this "beep" notification sound in the meeting room.

[0135] In some embodiments of this application, by performing echo path prediction on the audio signal emitted by the speaker of an electronic device, a predicted echo audio signal can be obtained. In this way, an audio signal corresponding to the predicted echo audio signal can be found in the first audio signal, and thus the echo audio signal can be obtained.

[0136] In some embodiments of this application, the denoising process performed on the sixth audio signal to obtain the seventh audio signal may specifically include: Extract the acoustic features of the sixth audio signal; Based on the acoustic characteristics of the sixth audio signal, determine the ambient noise audio signal; The environmental noise audio signal in the sixth audio signal is removed to obtain the seventh audio signal.

[0137] In some embodiments of this application, acoustic features of the sixth audio signal can be extracted. For example, the sixth audio signal can be processed using an acoustic recognition model to extract its acoustic features. Since the acoustic features of noise are different from those of human voice, the presence of environmental noise audio signals in the sixth audio signal can be determined based on these acoustic features.

[0138] Continuing with the example above, after eliminating the echo of the "beep" notification sound from the phone's speaker and the sound of the user typing on the keyboard, a sixth audio signal can be obtained. Then, acoustic features can be extracted from the sixth audio signal. Based on these acoustic features, it can be determined that there is air conditioner humming noise in the sixth audio signal, and the air conditioner humming noise can be removed from the sixth audio signal.

[0139] In the embodiments of this application, environmental noise audio signals are removed based on the acoustic characteristics of the audio signals. Since the noise reduction based on acoustic characteristics can more accurately distinguish between the noise spectrum and the speech spectrum, the complete frequency band and naturalness of the human voice can be preserved to the maximum extent while eliminating noise, thus avoiding the elimination of the human voice and achieving precise elimination of environmental noise.

[0140] In some embodiments of this application, to improve the accuracy of generating the first schedule information, the step of generating the first schedule information based on the second audio signal may specifically include: Text recognition is performed on the second audio signal to obtain the audio text information corresponding to the second audio signal; Acquire device placement status information, device display status information, device location information, and initial information; The audio text information, device placement status information, device display status information, device location information, and the first information are input into the second semantic analysis model. Through the second semantic analysis model, the user intent confidence, the first context confidence, and the second context confidence are obtained. The second semantic analysis model is used to calculate the scenario gate value by weighting the confidence scores of the first and second scenarios. The product of the context gating value and the user intent confidence level is calculated using the second semantic analysis model to determine the schedule generation decision score. The second semantic analysis model generates a decision score based on the schedule, and then generates the first instruction. If the first instruction instructs the generation of the first schedule, the first schedule information is generated based on the first text information.

[0141] The audio text information can be the text information obtained after performing text recognition on the second audio signal. For example, in the first example above, performing text recognition on the second audio signal "Come to my company for a meeting next Tuesday afternoon. Please bring the contract. What time is it? 3 PM. Do you need to bring any other materials? Also, please bring samples." yields the audio text information "Come to my company for a meeting next Tuesday afternoon. Please bring the contract. What time is it? 3 PM. Do you need to bring any other materials? Also, please bring samples."

[0142] The device placement state can be the placement state of electronic devices. The device placement state mentioned above can include the flat state, that is, the electronic device is placed flat. For example, in the first example above, the user places the mobile phone flat on the table after entering the meeting room, then the device placement state is the flat state.

[0143] Device display status information can refer to the status of the electronic device's display screen, such as whether the screen is off, on, or locked. The aforementioned device display status information can include whether the screen is locked or off.

[0144] Device location information can be the location information of an electronic device, which can be obtained based on the positioning module of the electronic device. In other words, the aforementioned device location information can include the geographical location of the electronic device.

[0145] The first piece of information could be used to indicate whether the user's current scheduled activity is a previously created activity. For example, in the first example above, the user's current scheduled activity is a meeting with Project Manager D and Sales Manager B; this activity is not among the user's previously created activities. In the second example above, the user's current scheduled activity is attending a product communication meeting, and this meeting's schedule is pre-recorded in the user's mobile calendar app; therefore, the user's current scheduled activity is among the created activities.

[0146] The second semantic analysis model can be a pre-trained model used to analyze audio text information, device placement status information, device display status information, device location information, and the first information. This second semantic analysis model can include, but is not limited to, deep learning-based neural network models, support vector machine models, or decision tree models. The specific type of the second semantic model can be selected according to user needs and is not limited in this embodiment. It should be noted that the second semantic model can be an LLM model deployed on the NPU. By deploying the first and second semantic models on different processing cores of the NPU and processing the information based on these different cores, extreme energy efficiency is achieved.

[0147] User intent confidence can be used to indicate the probability of recognizing the first text information in the audio text information. The aforementioned first text information can be the text information of the first schedule time and the first schedule item contained in the audio text information, that is, the first text information can include the first schedule time and the first schedule item. For example, if the audio text information in the first example above is "Come to my company for a meeting next Tuesday afternoon. Please bring the contract. What time is it? 3 PM. Do you need to bring any other materials? Also, please bring a sample.", then the first text information is the text information indicating the first schedule time "Next Tuesday afternoon at 3 PM" and the text information indicating the first schedule item "Come to my company for a meeting. Please bring the contract and a sample."

[0148] The first scenario confidence level can indicate the probability that the device is placed flat. The second scenario confidence level can indicate the probability that the first condition is met. The first condition can be that the device display status is off or in a closed state, the device location information is a reference geographical location, and the first information indicates that the user's currently executing schedule is an already created schedule. The reference geographical location can be a geographical location where the user needs to create a schedule, but it is inconvenient for the user to record the entire process or immediately stop their work to create the schedule. For example, the reference geographical location could be the user's company, office building, exhibition hall, etc.

[0149] The scenario gating value can be a value obtained by weighting the confidence levels of the first scenario and the confidence levels of the second scenario.

[0150] The schedule generation decision score can be the decision score for determining whether to generate the first schedule information. Specifically, the schedule generation decision score can be determined based on the product of the user intent confidence and the context gating value.

[0151] The first instruction can be used to indicate whether to generate a first schedule. Specifically, if the schedule generation decision score is greater than a second threshold, the first instruction instructs to generate the first schedule; if the schedule generation decision score is less than or equal to the second threshold, the first instruction instructs to skip generating the first schedule. The aforementioned second threshold can be a pre-set threshold for the schedule generation decision score, and the value of the second threshold is in the range of 0.5 to 1. For example, the second threshold can be 0.6, 0.8, or 0.9, etc. The specific value of the second threshold can be set by the user according to their needs, and is not limited in this embodiment.

[0152] In some embodiments of this application, text recognition can be performed on the second audio signal first to obtain the audio text information of the second audio signal. Then, device placement status information, device display status information, device location information and first information can be obtained. Then, the audio text information, device placement status information, device display status information, device location information and first information can be input into the second semantic analysis model. Through the second semantic analysis model, the user intent confidence, the first context confidence and the second context confidence can be obtained.

[0153] Specifically, using the second semantic analysis model, according to the following formula (1), the user intent confidence level of identifying the first text information containing the first schedule time and the first schedule item from the audio text information can be determined.

[0154] In the above formula (1), For audio text information, Based on user intent confidence. In the above formula (1) The range of values ​​is 0 < <1.

[0155] Then, using the second semantic analysis model, based on the device placement status information, device display status information, device location information, and the first information, the confidence of the first scenario and the confidence of the second scenario can be weighted and calculated according to the following formula (2) to obtain the scenario gate value.

[0156] In the above formula (2), The confidence level for the first scenario is... for The weight, For the second scenario confidence level, for The weight, The Sigmoid function is used to... Compressed to between 0 and 1, here and The settings can be customized according to user needs, and are not limited in this embodiment. In the above formula (2) It is a local constant used to avoid It is 0.

[0157] By using the second semantic analysis model, the product of the user intent confidence score and the context gating value is calculated to obtain the schedule generation decision score. If the schedule generation decision score is greater than the second threshold, a first instruction to indicate the generation of the first schedule can be generated. If the schedule generation decision score is less than or equal to the second threshold, a first instruction to indicate skipping the generation of the first schedule can be generated. If the second semantic analysis model generates a first instruction to indicate the generation of the first schedule, the first schedule information can be generated based on the first text information.

[0158] Continuing with the first example above, taking the audio text message "Come to my company for a meeting next Tuesday afternoon. Please bring your contract. What time is it? 3 PM. Do you need to bring any other materials? Also, please bring a sample." with a second threshold of 0.6, the second semantic analysis model identifies the first scheduled time "next Tuesday afternoon at 3 PM" and the first scheduled item "Come to my company for a meeting. Please bring your contract and a sample." The user intent confidence is 0.9. Then, the second semantic analysis model determines that the user's phone is placed flat, resulting in a first scenario confidence of 1. The second semantic analysis model also determines that the user's phone screen is off, the user is at their own company, and the currently executed schedule is not recorded in the phone's calendar app, resulting in a second scenario confidence of 0.67. Then, using the second semantic analysis model, the confidence of the first scenario and the confidence of the second scenario are weighted according to the above formula (2). If the obtained scenario gate value is 0.8, then the scenario gate value of 0.8 is greater than the second threshold of 0.6, and the first instruction can be generated. The first instruction indicates the generation of the first schedule. At this time, the first text information "Come to my company for a meeting next Tuesday at 3 pm. Please bring the contract and the sample" can be used to generate the first schedule. Figure 2 The first schedule information 211 is shown.

[0159] Continuing with the second example above, taking the audio text information "This phone will be available for sale at 00:00 on December 31, 2025" and the second threshold of 0.6 as an example, the second semantic analysis model identifies the first schedule time "00:00 on December 31, 2025" and the first schedule item "new phone available for sale" in the audio text information with a user intent confidence level of 0.9. Then, the second semantic analysis model determines that the user's phone is placed flat, resulting in a first scenario confidence level of 1. The second semantic analysis model also determines that the user's phone screen is off, the user is at an exhibition venue, and the user's currently executed schedule is recorded in the phone's calendar app, resulting in a second scenario confidence level of 1. Then, using the second semantic analysis model, the confidence of the first scenario and the confidence of the second scenario are weighted according to the above formula (2). If the obtained scenario gate value is 1, then the scenario gate value 1 is greater than the second threshold 0.6, and the first instruction can be generated. The first instruction indicates the generation of the first schedule. At this time, the first text information "This mobile phone will be available for sale at 0:00 on December 31, 2025" can be used to generate the first schedule. Figure 3 The first schedule information 211 is shown.

[0160] It should be noted that the product of user intent confidence and context gating value is used as the decision score for generating the first schedule information to give the context gating value a veto power over user intent confidence. For example, if Sales Manager B's phone is not placed flat on the table but in his pocket, it indicates that Sales Manager B did not intend to use his phone to record the meeting content and create a schedule. Therefore, even if Project Manager D tells Sales Manager B to come to his company next Tuesday at 3 pm with the contract and samples, the final decision score for generating the first schedule information will be close to 0, and the first schedule information will not be generated, thus preventing false alarms.

[0161] In the embodiments of this application, the generation of first schedule information is determined by simultaneously considering the user intent confidence level of whether the first text information containing the first schedule time and first schedule item is identified in the audio text information, as well as physical scenarios such as device placement status information, device display status information, device location information, and whether the user's currently executed schedule is within an already created schedule. That is, the generation of first schedule information is determined by a deep combination of physical scenarios and user intent, which achieves accurate identification of the user intent to create the schedule in the first audio signal, improves the accuracy of first schedule information generation, and reduces the false alarm rate. Furthermore, in the embodiments of this application, a second semantic analysis model is used to process the audio text information, device placement status information, device display status information, device location information, and whether the user's currently executed schedule is within an already created schedule to obtain a schedule generation decision score. Then, a first instruction indicating whether to generate the first schedule is directly generated based on the schedule generation decision score, instead of manually analyzing the audio text information, device placement status information, device display status information, device location information, and whether the user's currently executed schedule is within an already created schedule to determine whether to generate the first schedule. This improves the generation efficiency and objectivity of the first instruction indicating whether to generate the first schedule.

[0162] In some embodiments of this application, the audio text information may include a first schedule time. For example, in the first example above, the audio text information includes the first schedule time "Next Tuesday at 3 PM". The first window may also include a first control for determining the creation of the first schedule. This first control is as follows: Figure 2 The "OK" control 22 and Figure 3 The "OK" control 32 in the middle.

[0163] The methods mentioned above may also include: If the time of the first schedule is the same as the time of the second schedule of the second schedule, the second schedule information of the second schedule will be displayed in the first window; Receive the first input to the first control; In response to the first input, create the first schedule and delete the second schedule.

[0164] The second schedule can be a schedule already recorded in the schedule application of an electronic device. For example, in the above example, the user's mobile phone records the schedule of the product communication meeting that the user is currently attending, and the schedule of attending the group meeting at 3 pm on December 2, 2025. Then the schedules "attending the product communication meeting at 2 pm on November 30, 2025" and "attending the group meeting at 3 pm on December 2, 2025" are both second schedules.

[0165] The second schedule time refers to the schedule time within the second schedule. For example, in the second example above, for the second schedule "Attend the product communication meeting at 2 PM on November 30, 2025", the second schedule time is 2 PM on November 30, 2025. For the second schedule "Attend the team meeting at 3 PM on December 2, 2025", the second schedule time is 3 PM on December 2, 2025.

[0166] The second schedule information can be the schedule information for the second schedule, as described above. Figure 2 The second schedule information is 215.

[0167] The first input can be user input to the first control, used to create and delete a first schedule. Exemplarily, the first input includes, but is not limited to: touch input from the user using a finger or stylus, a voice command, a specific gesture, or other feasible input, which can be determined according to actual usage needs and is not limited in this embodiment. The specific gesture in this embodiment can be any one of a single-click gesture, a swipe gesture, a drag gesture, a pressure-recognition gesture, a long-press gesture, an area-change gesture, a double-press gesture, a double-tap gesture, or a shaking gesture; the click input in this embodiment can be a single-click input, a double-tap input, or any number of clicks, and can also be a long-press input or a short-press input. For example, the first input can be user click input to the first control.

[0168] In some embodiments of this application, when the first schedule time is the same as the second schedule time of the second schedule, the second schedule information of the second schedule can be displayed in the first window, and then in response to the first input to the first control, the first schedule can be created and the second schedule can be deleted.

[0169] Continuing with the first example above, the first schedule time is 3 PM next Tuesday, which is 3 PM on December 2nd, 2025. This first schedule time is the same as the second schedule time of 3 PM on December 2nd, 2025. Therefore, it can be done as follows: Figure 2 As shown, the second schedule information 215 is displayed in the first window 21. If the user clicks the "OK" control 22, the second schedule "Attend a group meeting at 3 pm on December 2, 2025" recorded in the phone's schedule application can be deleted, and the first schedule "Attend a meeting at Project Manager D's company next Tuesday at 3 pm with the contract and samples" can be created.

[0170] It should be noted that if the time of the first schedule coincides with the time of the second schedule, a schedule conflict warning message can be displayed in the first window, such as... Figure 2 The "Schedule Conflict" message 25 in the system.

[0171] In the embodiments of this application, if the first schedule and the second schedule conflict, and the user chooses to create the first schedule, the first schedule can be created and the second schedule can be deleted, thus avoiding the duplicate recording of schedules with the same time.

[0172] In some embodiments of this application, displaying the first window may specifically include: The first window is displayed on the lock screen of the electronic device.

[0173] In some embodiments of this application, when the display screen of the electronic device is in a screen-off state, the first window can be displayed on the lock screen interface of the electronic device.

[0174] In the embodiments of this application, by displaying a first window on the lock screen of an electronic device, the first schedule information can be pushed to the user on the lock screen through a silent interactive interface. This allows the user to be prompted to create the first schedule directly without unlocking the screen, even when the electronic device's display is off. This does not interrupt the user's current operation and only requires one click to complete the confirmation. The whole process is zero-disruption, low-power, and protects the security of the user's privacy data.

[0175] In some embodiments of this application, the first window described above may further include a first control and a second control. The second control here is as follows: Figure 2 The "ignore" control 23 and Figure 3 The "Ignore" control 33. This second control is used to skip the creation of the first schedule. That is, when the user clicks... Figure 2 When the "Ignore" control 23 is selected, the first schedule 214, "Go to Project Manager D's company for a meeting next Tuesday at 3 PM with the contract and samples," can be omitted. Figure 3 When using the "Ignore" control 33, the first schedule 314 for "New mobile phone to go on sale at 0:00 on December 31, 2025" can be omitted.

[0176] After displaying the first window, the method described above may further include: Upon receiving input to the first control, the reward value of the first semantic analysis model is determined to be the first reward value; based on the first reward value and the first probability, the model parameters of the first semantic analysis model are updated. Upon receiving input to the second control, the reward value of the first semantic analysis model is determined to be the second reward value; the model parameters of the first semantic analysis model are updated based on the first reward value and the second probability.

[0177] The first reward value is the reward for updating the model parameters in the first semantic analysis model when input is received to the first control. The second reward value is the reward for updating the model parameters in the first semantic analysis model when input is received to the second control. The first reward value is greater than the second reward value.

[0178] The first probability is the probability of creating the first schedule based on the first audio signal. The second probability is the probability of skipping the creation of the first schedule based on the first audio signal.

[0179] In some embodiments of this application, upon receiving input to the first control, the reward value of the first semantic analysis model can be determined as a first reward value, and then the model parameters of the first semantic analysis model can be updated based on the first reward value and the first probability.

[0180] Continuing with the first example above, if Sales Manager B clicks... Figure 2 The "Confirm" control 22 in the middle determines the reward value for rewarding the first semantic analysis model as the first reward value. Then, the model parameters of the first semantic analysis model can be updated based on the first reward value and the first probability of creating the first schedule 214 based on the first audio signal. This can enhance the sensitivity of the first semantic analysis model to specific voiceprints, specific terms, etc. in the audio signal.

[0181] Correspondingly, upon receiving input to the second control, the reward value of the first semantic analysis model is determined to be the second reward value; based on the first reward value and the second probability, the model parameters of the first semantic analysis model are updated.

[0182] Continuing with the first example above, if the user clicks... Figure 2 The "ignore" control 23 in the middle determines the reward value for rewarding the first semantic analysis model as the second reward value. Then, based on the second reward value and the probability of skipping the creation of the first schedule 214 based on the first audio signal, the model parameters of the first semantic analysis model can be updated. This can reduce the sensitivity of the first semantic analysis model to specific voiceprints, specific terms, etc. in the audio signal.

[0183] In some embodiments of this application, events are created based on reward values ​​and schedules, and the model parameters of the first semantic analysis model can be updated with reference to the following formula (3).

[0184] In the above formula (3), These are the model parameters for the first semantic analysis model. The updated model parameters for the first semantic analysis model. For state, In the state The following actions were taken. In the state The following actions The probability, i.e., the first probability or the second probability. The learning rate is determined based on the analysis accuracy of the first semantic analysis model. The specific value can be selected by the user according to their needs, and is not limited in this embodiment. This is the reward value, either the first reward value or the second reward value.

[0185] Continuing with the first example above, To receive the first audio signal, To create the first schedule 214, or, To receive the first audio signal, To skip creating the first schedule 214.

[0186] In the embodiments of this application, by responding to different user inputs to the first control and the second control, different reward values ​​for updating the model parameters of the first semantic analysis model can be determined, and then the model parameters of the first semantic analysis model can be updated according to the reward values. In this way, the model parameters of the first semantic analysis model can be updated differently according to the different choices of the user to the first control and the second control, so that the processing results of the first semantic analysis model are more in line with the user's personalized usage habits.

[0187] It should be noted that the information processing methods provided in the above embodiments are all completed in electronic devices. The audio signal is destroyed after use, and the audio signal is not stored or uploaded to the cloud, thus ensuring the security of the user's privacy data.

[0188] To better understand the information processing method provided in the embodiments of this application, please refer to... Figure 4 The information processing method provided in this application embodiment may include steps 41-54.

[0189] Step 41: Determine whether the electronic device is in a flat position, whether the display screen of the electronic device is in a screen-off state, whether a human voice signal is detected within a first time period, and whether the first probability is greater than the third threshold. If yes, proceed to step 42; otherwise, end the process.

[0190] In step 41, the method of determining whether the electronic device is in a flat position, whether the display screen of the electronic device is in a screen-off state, whether a human voice signal is detected within a first time period, and whether the first probability is greater than the third threshold can refer to the process of determining whether the electronic device is in a flat position, whether the display screen of the electronic device is in a screen-off state, whether a human voice signal is detected within a first time period, and whether the first probability is greater than the third threshold in the above embodiment, and will not be repeated here.

[0191] Step 42: Obtain the first audio signal.

[0192] Step 42 is the same as step 110 in the above embodiment, and will not be described again here.

[0193] It should be noted that after step 42, the first audio signal can also be subjected to noise reduction and enhancement processing. The specific process of noise reduction and enhancement processing of the first audio signal can be referred to in the above embodiment, which involves performing echo cancellation processing on the first audio signal to obtain the fifth audio signal; locating the sound source of the fifth audio signal according to the position of the microphone in the electronic device to obtain the sixth audio signal; and performing noise reduction processing on the sixth audio signal to obtain the seventh audio signal. The process will not be repeated here.

[0194] Step 43: Perform semantic recognition on the first audio signal to obtain the first recognition result.

[0195] Step 44: If the first recognition result indicates that the user intent to create a schedule is recognized based on the first audio signal, extract the audio signal containing the user intent to create a schedule from the first audio signal to obtain the second audio signal.

[0196] The process of steps 43-44 above is consistent with the process of performing semantic recognition on the first audio signal to obtain the first recognition result in the above embodiment; and extracting the audio signal containing the user intention to create the schedule from the first audio signal to obtain the second audio signal when the first recognition result indicates that the user intention to create the schedule is recognized based on the first audio signal. It will not be described again here.

[0197] Step 45: Perform text recognition on the second audio signal to obtain the audio text information corresponding to the second audio signal.

[0198] Step 46: Obtain device placement status information, device display status information, device location information, and first information.

[0199] In step 46, the device placement status information includes a flat position; the device display status information includes a locked screen or a screen-off state; the device location information includes the geographical location of the electronic device; and the first information is used to indicate whether the user's current execution schedule is a schedule that has already been created.

[0200] Step 47: Input the audio text information, device placement status information, device display status information, device location information, and the first information into the second semantic analysis model. Through the second semantic analysis model, obtain the user intent confidence, the first context confidence, and the second context confidence.

[0201] In step 47, the user intent confidence score is used to indicate the probability of recognizing the first text information in the audio text information; the first text information includes the first schedule time and the first schedule item; the first scenario confidence score is used to indicate the probability that the device placement status information is in a flat position; the second scenario confidence score is used to indicate the probability that the first condition is met; wherein, the first condition is that the device display status information is in a locked screen state or a screen-off state, the device location information is a reference geographical location, and the first information indicates that the user's currently executing schedule is a schedule that has already been created.

[0202] Step 48: Using the second semantic analysis model, the confidence scores of the first and second scenarios are weighted and calculated to obtain the scenario gating value.

[0203] Step 49: Calculate the product of the context gating value and the user intent confidence level using the second semantic analysis model to determine the schedule generation decision score.

[0204] Step 50: Using the second semantic analysis model, generate a decision score based on the schedule and generate the first instruction.

[0205] In step 50, the first instruction is used to indicate whether to generate the first schedule; if the schedule generation decision score is greater than the second threshold, the first instruction indicates to generate the first schedule; if the schedule generation decision score is less than or equal to the second threshold, the first instruction indicates to skip generating the first schedule.

[0206] Step 51: If the first instruction instructs to generate the first schedule, generate the first schedule information based on the first text information.

[0207] Step 52: Display the first window.

[0208] In step 52, the first window includes first schedule information, first calendar information, and first prompt information; the first schedule information includes descriptive information of the schedule elements of the first schedule; the schedule content of the first schedule is determined based on human voice signals; the first prompt information is used to prompt whether to create the first schedule; the first calendar information includes the first schedule time; the first schedule time is determined based on human voice signals.

[0209] It should be noted that the first window also includes a first control and a second control. The first control is used to indicate the creation of the first schedule, and the second control is used to indicate skipping the creation of the first schedule.

[0210] Step 53: Upon receiving input to the first control, determine the reward value of the first semantic analysis model as the first reward value, and update the model parameters of the first semantic analysis model based on the first reward value and the first probability.

[0211] The first probability mentioned above is the probability of creating the first schedule based on the first audio signal.

[0212] Step 54: Upon receiving input to the second control, determine the reward value of the first semantic analysis model as the second reward value, and update the model parameters of the first semantic analysis model based on the second reward value and the second probability.

[0213] The second probability mentioned above is the probability of skipping the creation of the first schedule based on the first audio signal. The first reward value in step 53 is greater than the second reward value in step 54.

[0214] The information processing method provided in this application can be executed by an information processing device. This application uses an information processing device executing the information processing method as an example to illustrate the information processing device provided in this application.

[0215] Figure 5 This is a schematic diagram illustrating the structure of an information processing apparatus according to an exemplary embodiment. For example... Figure 5 As shown, the information processing device 500 may include: Acquisition module 510 is used to acquire a first audio signal, wherein the first audio signal includes a human voice signal; The display module 520 is used to display a first window when a user intent to create a schedule is recognized based on the first audio signal; wherein the first window includes first schedule information, first calendar information and first prompt information; The first schedule information includes descriptive information of the schedule elements of the first schedule; the schedule content of the first schedule is determined based on the human voice signal; the first prompt information is used to prompt whether to create the first schedule; the first calendar information includes the first schedule time; the first schedule time is determined based on the human voice signal.

[0216] In the embodiments of this application, when the user's intention to create a schedule is identified in the first audio signal containing human voice signals, a first window can be displayed. This first window includes first schedule information, first calendar information, and a first prompt message. The first schedule information includes descriptive information of the schedule elements of the first schedule, and the schedule content of the first schedule is determined based on the human voice signal. The first prompt message is used to prompt whether to create the first schedule. The first calendar information includes the first schedule time; the first schedule time is determined based on the human voice signal. In this way, the electronic device can automatically and accurately extract core elements such as time and events from the schedule from continuous human voice signals and automatically fill them into the first window. Users do not need to listen to the entire recording on the electronic device or stop their current tasks to create the schedule. This eliminates the need for users to manually open the calendar application, input text one by one, and select dates, saving them a series of tedious operations. Users only need to confirm the first schedule information generated in the first window, simplifying multiple steps into a single confirmation. This greatly shortens the path and time for schedule creation, improves the accuracy and efficiency of schedule creation, and avoids interruptions to the user's current tasks in serious, immersive scenarios such as business meetings due to the need for manual schedule creation.

[0217] In some embodiments of this application, the apparatus further includes: The recognition module is used to perform semantic recognition on the first audio signal after acquiring the first audio signal to obtain a first recognition result; the first recognition result is used to indicate whether the user intent to create a schedule is recognized based on the first audio signal. The interception module is used to, when the first recognition result indicates that a user intent to create a schedule is recognized based on the first audio signal, intercept an audio signal containing the user intent to create a schedule from the first audio signal to obtain a second audio signal; The generation module is used to generate first schedule information based on the second audio signal.

[0218] In some embodiments of this application, the identification module is specifically used for: Semantic recognition is performed on the first audio signal to obtain the time intent confidence score and the event intent confidence score; Wherein, the time intent confidence is the probability of identifying the first schedule time from the first audio signal; the event intent confidence is the probability of identifying the first schedule event from the first audio signal; the first schedule information includes the first schedule event and the first schedule time; The device further includes: The determination module is configured to determine, when the maximum value of the time intent confidence score and the event intent confidence score is greater than a first threshold, a first recognition result indication that a user intent to create a schedule was identified based on the first audio signal.

[0219] In some embodiments of this application, the interception module is specifically used for: A third audio signal is extracted from the first audio signal; the third audio signal is an audio signal that includes the first schedule time and the first schedule item. From the first audio signal, identify second information associated with the first information; wherein the first information includes at least one of the following: the first schedule time; the first schedule item; Extract the fourth audio signal corresponding to the second information from the first audio signal; The third audio signal and the fourth audio signal are combined to obtain the second audio signal.

[0220] In some embodiments of this application, the identification module is specifically used for: The first audio signal is input into the first semantic analysis model. The first semantic analysis model is used to perform semantic recognition on the first audio signal to obtain the confidence level of time intent and the confidence level of event intent.

[0221] In some embodiments of this application, the apparatus further includes: The signal removal module is used to perform echo cancellation processing on the first audio signal after acquiring the first audio signal to obtain the fifth audio signal; The sound source localization module is used to locate the sound source of the fifth audio signal based on the position of the microphone in the electronic device, so as to obtain the sixth audio signal; The signal removal module is also used to perform noise reduction processing on the sixth audio signal to obtain the seventh audio signal; The identification module is specifically used for: Semantic recognition is performed on the seventh audio signal to obtain a first recognition result.

[0222] In some embodiments of this application, the electronic device includes N microphones; N is an integer greater than 1; The sound source localization module is specifically used for: Based on the time difference between each pair of microphones receiving the fifth audio signal in the electronic device, M time differences are obtained; ; Based on the M time differences, the fifth audio signal received by the N microphones is weighted and calculated to obtain the sixth audio signal.

[0223] In some embodiments of this application, the generation module is specifically used for: Perform text recognition on the second audio signal to obtain the audio text information corresponding to the second audio signal; The system acquires device placement status information, device display status information, device location information, and first information; wherein, the device placement status information includes a flat position; the device display status information includes a locked screen or a screen-off state; the device location information includes the geographical location of the electronic device; and the first information is used to indicate whether the user's currently executing schedule is a previously created schedule. The audio text information, device placement status information, device display status information, device location information, and first information are input into a second semantic analysis model. Through the second semantic analysis model, user intent confidence, first scenario confidence, and second scenario confidence are obtained. The user intent confidence indicates the probability of recognizing the first text information in the audio text information. The first text information includes the first schedule time and the first schedule item. The first scenario confidence indicates the probability that the device placement status is in a flat position. The second scenario confidence indicates the probability that a first condition is met. The first condition is that the device display status is locked or off, the device location information is a reference geographical location, and the first information indicates that the user's currently executing schedule is a created schedule. The first scenario confidence score and the second scenario confidence score are weighted and calculated using the second semantic analysis model to obtain the scenario gating value. The product of the context gating value and the user intent confidence level is calculated using the second semantic analysis model to determine the schedule generation decision score. Using the second semantic analysis model, a first instruction is generated based on the schedule generation decision score; the first instruction is used to indicate whether to generate the first schedule; if the schedule generation decision score is greater than a second threshold, the first instruction indicates to generate the first schedule; if the schedule generation decision score is less than or equal to the second threshold, the first instruction indicates to skip generating the first schedule. When the first instruction instructs the generation of a first schedule, first schedule information is generated based on the first text information.

[0224] In some embodiments of this application, the audio text information includes a first schedule time; the first window further includes a first control; the first control is used to determine the creation of the first schedule; The display module is further configured to: display the second schedule information of the second schedule in the first window when the first schedule time is the same as the second schedule time of the second schedule; The device further includes: A receiving module is used to receive the first input to the first control; A schedule processing module is used to create the first schedule and delete the second schedule in response to the first input.

[0225] In some embodiments of this application, the display module is specifically used for: The first window is displayed on the lock screen of the electronic device.

[0226] In some embodiments of this application, the first window further includes a first control and a second control; the second control is used to skip the creation of the first schedule; The device further includes: The model parameter update module is configured to, after displaying the first window and upon receiving input to the first control, determine that the reward value of the first semantic analysis model is a first reward value; update the model parameters of the first semantic analysis model according to the first reward value and a first probability, wherein the first probability is the probability of creating the first schedule based on the first audio signal; upon receiving input to the second control, determine that the reward value of the first semantic analysis model is a second reward value; update the model parameters of the first semantic analysis model according to the second reward value and a second probability, wherein the second probability is the probability of skipping the creation of the first schedule based on the first audio signal; wherein the first reward value is greater than the second reward value.

[0227] In some embodiments of this application, the acquisition module is specifically used for: The first audio signal is acquired when the electronic device is placed flat, the screen of the electronic device is off, a human voice signal is detected within the first time period, and the third probability is greater than the third threshold. The third probability is used to indicate the probability of detecting a human voice signal within the first time period.

[0228] The information processing device in this application embodiment can be an electronic device or a component within an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices besides a terminal. For example, the electronic device can be a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), television set (TV), ATM, or self-service machine, etc. This application embodiment does not specifically limit the device.

[0229] The information processing device in this application embodiment can be a device with an operating system. The operating system can be Android, iOS, or other possible operating systems; this application embodiment does not specifically limit the specific operating system.

[0230] The information processing device provided in this application embodiment can achieve... Figure 3 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0231] Optionally, such as Figure 6 As shown, this application embodiment also provides an electronic device 600, including a processor 601 and a memory 602. The memory 602 stores a program or instructions that can run on the processor 601. When the program or instructions are executed by the processor 601, they implement the various steps of the above-described information processing method embodiment and can achieve the same technical effect. To avoid repetition, they will not be described again here.

[0232] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0233] Figure 7 A schematic diagram of the hardware structure of an electronic device to implement an embodiment of this application.

[0234] The electronic device 700 includes, but is not limited to, components such as: radio frequency unit 701, network module 702, audio output unit 703, input unit 704, sensor 705, display unit 706, user input unit 707, interface unit 708, memory 709, and processor 710.

[0235] Those skilled in the art will understand that the electronic device 700 may also include a power supply (such as a battery) for supplying power to various components. The power supply may be logically connected to the processor 710 through a power management system, thereby enabling functions such as managing charging, discharging, and power consumption through the power management system. Figure 7 The electronic device structure shown does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements, which will not be elaborated here.

[0236] The processor 710 is used to acquire a first audio signal, which includes a human voice signal. Display unit 706 is configured to display a first window when a user intent to create a schedule is recognized based on the first audio signal; wherein the first window includes first schedule information, first calendar information, and first prompt information; The first schedule information includes descriptive information of the schedule elements of the first schedule; the schedule content of the first schedule is determined based on the human voice signal; the first prompt information is used to prompt whether to create the first schedule; the first calendar information includes the first schedule time; the first schedule time is determined based on the human voice signal.

[0237] Thus, upon recognizing the user's intent to create a schedule from the first audio signal containing human voice signals, a first window can be displayed. This first window includes first schedule information, first calendar information, and a first prompt. The first schedule information includes descriptive information about the schedule elements, and the schedule content is determined based on the human voice signal. The first prompt prompts whether to create the first schedule. The first calendar information includes the first schedule time, which is also determined based on the human voice signal. In this way, the electronic device can automatically and accurately extract core elements such as time and events from the schedule from continuous human voice signals and automatically fill them into the first window. Users do not need to listen to the entire recording on the electronic device or stop their current tasks to create the schedule. This eliminates the need for users to manually open the calendar application, input text one by one, and select dates—a series of tedious operations. Users only need to confirm the first schedule information generated in the first window. This simplifies multiple steps into a single confirmation, greatly shortening the schedule creation path and time, improving the accuracy and efficiency of schedule creation, and preventing interruptions to the user's current tasks in serious, immersive scenarios such as business meetings due to the need to manually create a schedule.

[0238] Optionally, the processor 710 is further configured to perform semantic recognition on the first audio signal to obtain a first recognition result; the first recognition result is used to indicate whether a user intent to create a schedule is recognized based on the first audio signal; if the first recognition result indicates that a user intent to create a schedule is recognized based on the first audio signal, an audio signal containing the user intent to create a schedule is extracted from the first audio signal to obtain a second audio signal; and first schedule information is generated based on the second audio signal.

[0239] Optionally, the processor 710 is further configured to perform semantic recognition on the first audio signal to obtain a time intent confidence score and an event intent confidence score; wherein the time intent confidence score is the probability of recognizing a first schedule time from the first audio signal; the event intent confidence score is the probability of recognizing a first schedule event from the first audio signal; the first schedule information includes the first schedule event and the first schedule time; if the maximum value of the time intent confidence score and the event intent confidence score is greater than a first threshold, a first recognition result is determined to indicate that a schedule creation user intent is recognized based on the first audio signal.

[0240] Optionally, the processor 710 is further configured to extract a third audio signal from the first audio signal; the third audio signal being an audio signal including the first schedule time and the first schedule item; identify second information associated with the first information from the first audio signal; wherein the first information includes at least one of the following: the first schedule time; the first schedule item; extract a fourth audio signal corresponding to the second information from the first audio signal; and synthesize the third audio signal and the fourth audio signal to obtain a second audio signal.

[0241] Optionally, the processor 710 is further configured to input the first audio signal into a first semantic analysis model, and through the first semantic analysis model, perform semantic recognition on the first audio signal to obtain time intent confidence and event intent confidence.

[0242] Optionally, the processor 710 is further configured to perform echo cancellation processing on the first audio signal to obtain a fifth audio signal; perform sound source localization on the fifth audio signal according to the position of the microphone in the electronic device to obtain a sixth audio signal; perform noise reduction processing on the sixth audio signal to obtain a seventh audio signal; and perform semantic recognition on the seventh audio signal to obtain a first recognition result.

[0243] Optionally, the electronic device includes N microphones; N is an integer greater than 1; the processor 710 is further configured to obtain M time differences based on the time difference between each pair of microphones receiving the fifth audio signal in the electronic device; Based on the M time differences, the fifth audio signal received by the N microphones is weighted and calculated to obtain the sixth audio signal.

[0244] Optionally, the processor 710 is further configured to perform text recognition on the second audio signal to obtain audio text information corresponding to the second audio signal; acquire device placement status information, device display status information, device location information, and first information; wherein, the device placement status information includes a flat position; the device display status information includes a locked screen or a screen-off state; the device location information includes the geographical location of the electronic device; the first information is used to indicate whether the user's currently executing schedule is a created schedule; input the audio text information, the device placement status information, the device display status information, the device location information, and the first information into a second semantic analysis model, and obtain user intent confidence, first context confidence, and second context confidence through the second semantic analysis model; the user intent confidence is used to indicate the probability of recognizing the first text information in the audio text information; the first text information includes the first schedule time and the first schedule item; the first context confidence is used to indicate the probability that the device placement status is a flat position; The second scenario confidence score indicates the probability that the first condition is met; wherein, the first condition is that the device display status information is in a locked screen state or a screen-off state, the device location information is a reference geographical location, and the first information indicates that the user's currently executing schedule is a created schedule; the first scenario confidence score and the second scenario confidence score are weighted and calculated using the second semantic analysis model to obtain a scenario gating value; the product of the scenario gating value and the user intent confidence score is calculated using the second semantic analysis model to determine the schedule generation decision score; a first instruction is generated based on the schedule generation decision score using the second semantic analysis model; the first instruction indicates whether to generate the first schedule; if the schedule generation decision score is greater than a second threshold, the first instruction indicates to generate the first schedule; if the schedule generation decision score is less than or equal to the second threshold, the first instruction indicates to skip generating the first schedule; if the first instruction indicates to generate the first schedule, the first schedule information is generated based on the first text information.

[0245] Optionally, the audio text information includes a first schedule time; the first window further includes a first control; the first control is used to determine the creation of the first schedule; the display unit 706 is also used to display the second schedule information of the second schedule in the first window when the first schedule time is the same as the second schedule time of the second schedule. User input unit 707 is used to receive a first input to the first control; The processor 710 is also configured to, in response to the first input, create the first schedule and delete the second schedule.

[0246] Optionally, the display unit 706 is also used to display a first window on the lock screen interface of the electronic device.

[0247] Optionally, the first window further includes a first control and a second control; the second control is used to skip creating the first schedule; the processor 710 is further configured to, upon receiving input to the first control, determine that the reward value of the first semantic analysis model is a first reward value; update the model parameters of the first semantic analysis model according to the first reward value and a first probability, wherein the first probability is the probability of creating the first schedule based on the first audio signal; upon receiving input to the second control, determine that the reward value of the first semantic analysis model is a second reward value; update the model parameters of the first semantic analysis model according to the second reward value and a second probability, wherein the second probability is the probability of skipping the creation of the first schedule based on the first audio signal; wherein the first reward value is greater than the second reward value.

[0248] Optionally, the processor 710 is further configured to acquire a first audio signal when the electronic device is in a flat position, the display screen of the electronic device is in a screen-off state, a human voice signal is detected within a first duration, and a third probability is greater than a third threshold; wherein the third probability is used to indicate the probability of detecting a human voice signal within the first duration.

[0249] It should be understood that, in this embodiment, the input unit 704 may include a graphics processing unit (GPU) 7041 and a microphone 7042. The GPU 7041 processes image data of still images or videos obtained by an image capture device (such as a color camera) in video capture mode or image capture mode. The display unit 706 may include a display panel 7061, which may be configured in the form of a liquid crystal display, an organic light-emitting diode, or the like. The user input unit 707 includes at least one of a touch panel 7071 and other input devices 7072. The touch panel 7071 is also called a touch screen. The touch panel 7071 may include a touch detection device and a touch controller. Other input devices 7072 may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, power buttons, etc.), trackballs, mice, and joysticks, which will not be described in detail here.

[0250] The memory 709 can be used to store software programs and various data. The memory 709 may primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, the memory 709 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DRRAM). The memory 709 in the embodiments of this application includes, but is not limited to, these and any other suitable types of memory.

[0251] Processor 710 may include one or more processing units; optionally, processor 710 integrates an application processor and a modem processor, wherein the application processor mainly handles operations involving the operating system, user interface, and applications, and the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into processor 710.

[0252] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described information processing method embodiments and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0253] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0254] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above information processing method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0255] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0256] This application provides a computer program product, which is stored in a storage medium and executed by at least one processor to implement the various processes of the information processing method embodiments described above, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0257] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0258] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0259] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. An information processing method, characterized in that, The method includes: Acquire a first audio signal, wherein the first audio signal includes a human voice signal; If a user intent to create a schedule is recognized based on the first audio signal, a first window is displayed; wherein the first window includes first schedule information, first calendar information, and first prompt information; The first schedule information includes descriptive information of the schedule elements of the first schedule; the schedule content of the first schedule is determined based on the human voice signal; the first prompt information is used to prompt whether to create the first schedule; the first calendar information includes the first schedule time; the first schedule time is determined based on the human voice signal.

2. The method according to claim 1, characterized in that, After acquiring the first audio signal, the method further includes: Semantic recognition is performed on the first audio signal to obtain a first recognition result; the first recognition result is used to indicate whether a user intent to create a schedule is recognized based on the first audio signal. If the first recognition result indicates that a user intent to create a schedule is recognized based on the first audio signal, an audio signal containing the user intent to create a schedule is extracted from the first audio signal to obtain a second audio signal; First schedule information is generated based on the second audio signal.

3. The method according to claim 2, characterized in that, The step of performing semantic recognition on the first audio signal to obtain a first recognition result includes: Semantic recognition is performed on the first audio signal to obtain the time intent confidence score and the event intent confidence score; Wherein, the time intent confidence is the probability of identifying the first schedule time from the first audio signal; the item intent confidence is the probability of identifying the first schedule item from the first audio signal; the first schedule information includes the first schedule item and the first schedule time; The method further includes: If the maximum value of the time intent confidence score and the event intent confidence score is greater than a first threshold, the first recognition result indicates that a user intent to create a schedule was identified based on the first audio signal.

4. The method according to claim 3, characterized in that, The step of extracting the audio signal containing the user's intent to create the schedule from the first audio signal to obtain the second audio signal includes: A third audio signal is extracted from the first audio signal; the third audio signal is an audio signal that includes the first schedule time and the first schedule item. From the first audio signal, identify second information associated with the first information; wherein the first information includes at least one of the following: the first schedule time; the first schedule item; Extract the fourth audio signal corresponding to the second information from the first audio signal; The third audio signal and the fourth audio signal are combined to obtain the second audio signal.

5. The method according to claim 3, characterized in that, The step of performing semantic recognition on the first audio signal to obtain time intent confidence and event intent confidence includes: The first audio signal is input into the first semantic analysis model. The first semantic analysis model is used to perform semantic recognition on the first audio signal to obtain the confidence level of time intent and the confidence level of event intent.

6. The method according to claim 2, characterized in that, After acquiring the first audio signal, the method further includes: The first audio signal is subjected to echo cancellation processing to obtain the fifth audio signal; Based on the position of the microphone in the electronic device, the sound source of the fifth audio signal is located to obtain the sixth audio signal; The sixth audio signal is denoised to obtain the seventh audio signal; The step of performing semantic recognition on the first audio signal to obtain a first recognition result includes: Semantic recognition is performed on the seventh audio signal to obtain a first recognition result.

7. The method according to claim 6, characterized in that, The electronic device includes N microphones; N is an integer greater than 1; The step of locating the sound source of the fifth audio signal based on the position of the microphone in the electronic device to obtain the sixth audio signal includes: Based on the time difference between each pair of microphones receiving the fifth audio signal in the electronic device, M time differences are obtained; ; Based on the M time differences, the fifth audio signal received by the N microphones is weighted and calculated to obtain the sixth audio signal.

8. The method according to claim 2, characterized in that, The step of generating the first schedule information based on the second audio signal includes: Perform text recognition on the second audio signal to obtain the audio text information corresponding to the second audio signal; The system acquires device placement status information, device display status information, device location information, and first information; wherein, the device placement status information includes a flat position; the device display status information includes a locked screen or a screen-off state; the device location information includes the geographical location of the electronic device; and the first information is used to indicate whether the user's currently executing schedule is a schedule that has already been created. The audio text information, device placement status information, device display status information, device location information, and first information are input into a second semantic analysis model. Through the second semantic analysis model, user intent confidence, first scenario confidence, and second scenario confidence are obtained. The user intent confidence indicates the probability of recognizing the first text information in the audio text information. The first text information includes the first schedule time and the first schedule item. The first scenario confidence indicates the probability that the device placement status is in a flat position. The second scenario confidence indicates the probability that a first condition is met. The first condition is that the device display status is locked or off, the device location information is a reference geographical location, and the first information indicates that the user's currently executing schedule is a created schedule. The first scenario confidence score and the second scenario confidence score are weighted and calculated using the second semantic analysis model to obtain the scenario gating value. The product of the context gating value and the user intent confidence level is calculated using the second semantic analysis model to determine the schedule generation decision score. Using the second semantic analysis model, a first instruction is generated based on the schedule generation decision score; the first instruction is used to indicate whether to generate the first schedule; if the schedule generation decision score is greater than a second threshold, the first instruction indicates to generate the first schedule; if the schedule generation decision score is less than or equal to the second threshold, the first instruction indicates to skip generating the first schedule. When the first instruction instructs the generation of a first schedule, first schedule information is generated based on the first text information.

9. The method according to claim 8, characterized in that, The audio text information includes a first schedule time; the first window also includes a first control; the first control is used to determine the creation of the first schedule; The method further includes: If the first schedule time is the same as the second schedule time of the second schedule, the second schedule information of the second schedule will be displayed in the first window; Receive the first input to the first control; In response to the first input, the first schedule is created, and the second schedule is deleted.

10. The method according to claim 1, characterized in that, The first window display includes: The first window is displayed on the lock screen of the electronic device.

11. The method according to claim 5, characterized in that, The first window also includes a first control and a second control; the second control is used to skip the creation of the first schedule. After displaying the first window, the method further includes: Upon receiving input to the first control, the reward value of the first semantic analysis model is determined to be a first reward value; the model parameters of the first semantic analysis model are updated according to the first reward value and a first probability; the first probability is the probability of creating the first schedule based on the first audio signal; Upon receiving input to the second control, the reward value of the first semantic analysis model is determined to be the second reward value; The model parameters of the first semantic analysis model are updated based on the second reward value and the second probability, where the second probability is the probability of skipping the creation of the first schedule based on the first audio signal. The first reward value is greater than the second reward value.

12. The method according to claim 1, characterized in that, The acquisition of the first audio signal includes: The first audio signal is acquired when the electronic device is placed flat, the screen of the electronic device is off, a human voice signal is detected within the first time period, and the third probability is greater than the third threshold. The third probability is used to indicate the probability of detecting a human voice signal within the first time period.

13. An information processing device, characterized in that, The device includes: An acquisition module is used to acquire a first audio signal, wherein the first audio signal includes a human voice signal; The display module is used to display a first window when the user's intention to create a schedule is recognized based on the first audio signal; wherein the first window includes first schedule information, first calendar information and first prompt information; The first schedule information includes descriptive information of the schedule elements of the first schedule; the schedule content of the first schedule is determined based on the human voice signal; the first prompt information is used to prompt whether to create the first schedule; the first calendar information includes the first schedule time; the first schedule time is determined based on the human voice signal.

14. The apparatus according to claim 13, characterized in that, The device further includes: The recognition module is used to perform semantic recognition on the first audio signal after acquiring the first audio signal to obtain a first recognition result; the first recognition result is used to indicate whether the user intent to create a schedule is recognized based on the first audio signal. The interception module is used to intercept an audio signal containing the user intent to create a schedule from the first audio signal when the first recognition result indicates that the user intent to create a schedule is recognized based on the first audio signal, so as to obtain a second audio signal; The generation module is used to generate first schedule information based on the second audio signal.

15. The apparatus according to claim 14, characterized in that, The identification module is specifically used for: Semantic recognition is performed on the first audio signal to obtain the time intent confidence score and the event intent confidence score; Wherein, the time intent confidence is the probability of identifying the first schedule time from the first audio signal; the item intent confidence is the probability of identifying the first schedule item from the first audio signal; the first schedule information includes the first schedule item and the first schedule time; The device further includes: The determination module is configured to determine, when the maximum value of the time intent confidence score and the event intent confidence score is greater than a first threshold, a first recognition result indication that a user intent to create a schedule was identified based on the first audio signal.

16. The apparatus according to claim 15, characterized in that, The interception module is specifically used for: A third audio signal is extracted from the first audio signal; the third audio signal is an audio signal that includes the first schedule time and the first schedule item. From the first audio signal, identify second information associated with the first information; wherein the first information includes at least one of the following: the first schedule time; the first schedule item; Extract the fourth audio signal corresponding to the second information from the first audio signal; The third audio signal and the fourth audio signal are combined to obtain the second audio signal.

17. The apparatus according to claim 15, characterized in that, The identification module is specifically used for: The first audio signal is input into the first semantic analysis model. The first semantic analysis model is used to perform semantic recognition on the first audio signal to obtain the confidence level of time intent and the confidence level of event intent.

18. The apparatus according to claim 14, characterized in that, The device further includes: The signal removal module is used to perform echo cancellation processing on the first audio signal after acquiring the first audio signal to obtain the fifth audio signal; The sound source localization module is used to locate the sound source of the fifth audio signal based on the position of the microphone in the electronic device, so as to obtain the sixth audio signal; The signal removal module is also used to perform noise reduction processing on the sixth audio signal to obtain the seventh audio signal; The identification module is specifically used for: Semantic recognition is performed on the seventh audio signal to obtain a first recognition result.

19. The apparatus according to claim 18, characterized in that, The electronic device includes N microphones; N is an integer greater than 1; The sound source localization module is specifically used for: Based on the time difference between each pair of microphones receiving the fifth audio signal in the electronic device, M time differences are obtained; ; Based on the M time differences, the fifth audio signal received by the N microphones is weighted and calculated to obtain the sixth audio signal.

20. An electronic device, characterized in that, It includes a processor and a memory, the memory storing a program or instructions that can run on the processor, the program or instructions being executed by the processor to implement the steps of the information processing method as described in any one of claims 1-12.

Images