Task processing method and device and electronic equipment

Abstract

The invention provides a task processing method and device and electronic equipment, and the method comprises the steps: obtaining a target task if the electronic equipment is in a first running state and a target application is in a running state; the target application can call the target model for responding to the target task; the target application responds to the target task and outputs an output result for the target task; wherein the output result corresponds to the security level corresponding to the electronic equipment in the first running state; the security level of the electronic equipment in the first running state is lower than that of the electronic equipment in the second running state.

Description

Technical Field

[0001] The present application relates to the field of generative large model applications, and more specifically to a task processing method, device and electronic device. Background Art

[0002] In recent years, with the changes, development and application of artificial intelligence (AI) technology, the functions of electronic devices such as mobile phones and laptops have become more comprehensive and intelligent. They can use artificial intelligence algorithms to automatically generate content that meets user intentions, such as calling AIGC (Artificial Intelligence Generated Content) models to automatically query and summarize hot news, create product promotion videos or chat with users, etc., which provides great convenience for users' lives and work.

[0003] In order to ensure the security of sensitive data of electronic devices and users, the above task processing method can only be executed when the electronic device is unlocked. Once the electronic device enters the lock screen state, the electronic device will be prohibited from outputting the above content that meets the user's intention, which cannot meet the user's task processing needs for the electronic device.

[0004] In view of the above problems, this application provides the following solutions:

[0005] In a first aspect, the present application proposes a task processing method, the method comprising:

[0006] If the electronic device is in a first running state and the target application is in a running state, obtaining a target task, the target application can call a target model to respond to the target task;

[0007] The target application outputs an output result for the target task in response to the target task;

[0008] Wherein, the output result corresponds to the security level corresponding to the electronic device being in the first operating state;

[0009] The security level of the electronic device in the first operating state is lower than the security level of the electronic device in the second operating state.

[0010] In a possible implementation, before outputting the output result for the target task, the method further includes:

[0011] Obtaining a processing result of the target model in response to the target task;

[0012] Process the processing result based on the first security policy to obtain the output result;

[0013] Wherein, the first security policy is a security policy corresponding to a security level corresponding to the electronic device being in a first operating state;

[0014] The semantics represented by the output result is the same as the semantics represented by the processing result, and the security of the output result is higher than the security of the processing result.

[0015] In a possible implementation, before the target model obtains the target data required to respond to the target task, the method further includes:

[0016] obtaining the target data;

[0017] The target data is processed based on a first security policy to obtain security data; wherein the first security policy is a security policy corresponding to a security level corresponding to the electronic device being in a first operating state, and the security of the security data is higher than the security of the target data.

[0018] The safety data is provided to the target model so that the target model responds to the target task based on the safety data.

[0019] In a possible implementation, the target application responding to the target task includes:

[0020] The target application configures the target model based on the first configuration parameter so that the target model responds to the target task based on the feature information of the public knowledge base;

[0021] Wherein, the first configuration parameter is a configuration parameter corresponding to the security level corresponding to the electronic device being in the first operating state;

[0022] The security of the output result generated by the target model in response to the target task based on the feature information of the public knowledge base is higher than the security of the output result generated by the target model in response to the target task based on the feature information of the personal knowledge base.

[0023] In a possible implementation, the target application outputs an output result for the target task in response to the target task, and the method also includes:

[0024] Get target parameters;

[0025] If the target parameter determines that the target condition is satisfied, executing the target application in response to the target task to output an output result for the target task;

[0026] The target parameter determination that meets the target condition represents that the process of obtaining the target task is in a trusted environment.

[0027] In a possible implementation, the first security policy includes at least one of the following:

[0028] A filtering strategy based on a sensitive word library is used to prohibit the output result from containing sensitive words in the sensitive word library;

[0029] An anonymization processing strategy for shielding user information and device information to prohibit the output result from including device information and user information of the electronic device;

[0030] Privacy protection strategy based on differential privacy technology.

[0031] In a possible implementation, the target parameter determines that the target condition is satisfied, including at least one of the following:

[0032] If the target parameter includes location information of the environment in which the electronic device is located, determining that the location information matches location information of any trusted environment stored by the electronic device;

[0033] If the target parameter includes a device identification of at least one output device communicatively connected to the electronic device, it is determined that the device identification matches a device identification of a trusted device bound to the electronic device.

[0034] In a possible implementation, before outputting the output result for the target task, the method further includes:

[0035] Performing security verification on the output result;

[0036] If the security check result is failed, the response to the target task is terminated to prohibit the output of the output result for the target task, or a security level switching prompt message is output to control the electronic device to switch from the first operating state to the second operating state in response to the operating state switching operation input by the user, so that the target model responds to the target task and outputs the output result for the target task corresponding to the security level corresponding to the electronic device being in the second operating state.

[0037] A second aspect of the present application provides a task processing device, the device comprising:

[0038] A target task obtaining module, used for obtaining a target task if the electronic device is in a first running state and the target application is in a running state, wherein the target application can call a target model to respond to the target task;

[0039] An output module, configured for the target application to output an output result for the target task in response to the target task;

[0040] Wherein, the output result corresponds to the security level corresponding to the electronic device being in the first operating state;

[0041] The security level of the electronic device in the first operating state is lower than the security level of the electronic device in the second operating state.

[0042] A third aspect of the present application provides an electronic device, the electronic device comprising: at least one memory and at least one processor, wherein:

[0043] A memory for storing computer program instructions and program codes for supporting the operation of a target application;

[0044] The processor is used to load and execute the computer program instructions to implement the following steps:

[0045] If the electronic device is in a first running state and the target application is in a running state, a target task is obtained, and the target application can call a target model to respond to the target task;

[0046] The target application outputs an output result for the target task in response to the target task;

[0047] Wherein, the output result corresponds to the security level corresponding to the electronic device being in the first operating state;

[0048] The security level of the electronic device in the first operating state is lower than the security level of the electronic device in the second operating state. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0050] Figure 1 A flowchart of a task processing method provided in Example 1 of the present application;

[0051] Figure 2 A flowchart of a task processing method provided in Example 2 of the present application;

[0052] Figure 3 A flowchart of a task processing method provided in Example 3 of the present application;

[0053] Figure 4 A flowchart of a task processing method provided in Embodiment 4 of the present application;

[0054] Figure 5A flowchart of a task processing method provided in Embodiment 5 of the present application;

[0055] Figure 6 A flowchart of a task processing method provided in Example 6 of the present application;

[0056] Figure 7 A flowchart of a task processing method provided in Embodiment 7 of the present application;

[0057] Figure 8 A flowchart of a task processing method provided in Embodiment 8 of the present application;

[0058] Fig. 9 A schematic diagram of the structure of a task processing device provided in Example 1 of the present application;

[0059] Fig.10 A schematic diagram of the structure of a task processing device provided in Embodiment 2 of the present application;

[0060] Fig.11 A schematic diagram of the structure of a task processing device provided in Embodiment 3 of the present application;

[0061] Fig.12 A schematic diagram of the structure of a task processing device provided in Embodiment 4 of the present application;

[0062] Fig.13 A schematic diagram of the hardware structure of an electronic device provided in Embodiment 1 of the present application;

[0063] Fig.14 A schematic diagram of the hardware structure of an electronic device provided in Example 2 of the present application. DETAILED DESCRIPTION

[0064] The embodiments of the present application are described below in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation mode of the present application are only used to explain the specific embodiments of the present application and are not intended to limit the present application. The embodiments of the present application are described below in conjunction with the drawings. It is known to those of ordinary skill in the art that with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0065] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, which is only to describe the distinction mode adopted by the objects of the same attributes when describing in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0066] In order to solve the above problems, the embodiment of the present application provides a task processing method. When the electronic device is in a low security level operating state, such as in a locked screen state, in order to meet the user's target task processing needs, the target application is allowed to call the target model to respond to the target task, but the output result of the electronic device for the target task is guaranteed to correspond to the lower security level, avoiding the direct output of the processing content for the target task (such as sensitive information of the user and the electronic device, or pre-defined important data content, etc.) that is prohibited from being output by the lower security level, so as to meet the user's target task processing needs while ensuring security and improve the user experience. The task processing method of the embodiment of the present application is described in detail below in conjunction with the accompanying drawings.

[0067] Reference Figure 1 , is a flowchart of a task processing method provided in the first embodiment of the present application, which can be applied to electronic devices, such as smart phones, tablet computers, wearable devices, augmented reality (AR) / virtual reality (VR) devices, laptop computers or other display devices, and any other user terminals in the form of products. Figure 1 As shown, the task processing method may include:

[0068] Step S11, if the electronic device is in the first running state and the target application is in the running state, obtaining a target task; the target application can call the target model to respond to the target task;

[0069] In the embodiments of the present application, with the changes and development of artificial intelligence, the target model can be a large AI model such as a Generative Artificial Intelligence (GAI) model or various AIGC (Artificial Intelligence Generated Content) models, or a large model (Large Model, LM), a large language model (Large Language Model, LLM), etc., or other machine learning / deep learning models trained for target tasks based on one or more artificial intelligence algorithms, etc. This application does not limit the type of target model and its acquisition method.

[0070] Preferably, the present application may use a large AI model such as the one described above as the target model, and the electronic device may be a user terminal configured with corresponding software and hardware resources that support the operation or call of the large AI model, so that when it is determined that the obtained target task needs to call the large AI model to implement, the resource can reliably support the target model to respond to the target task. It can be seen that the present application selects the large AI model as the target model for responding to the target task, which can make full use of the AI ​​model's ability to be more intelligent, comprehensive, and generate more accurate content, meet more task processing needs, and improve the diversity and accuracy of task processing. The implementation process of various AI large models responding to different types of target tasks in the embodiments of the present application is not described in detail, but the target model used by the electronic device is not limited to this large AI model.

[0071] Among them, when the target model is the above-mentioned AI large model, the target tasks that need to call the target model response in this application may include one or more content generation tasks such as creating text, images, videos, audio or software codes, such as product design drawing tasks, character, scene and storyline generation tasks for the game to be designed, document theme summary tasks, etc. This application does not limit the task content of the target task and the method of obtaining it. It should be understood that the target tasks obtained in this application include but are not limited to the content generation tasks described above, and may also include other question-and-answer tasks or control tasks, etc., which can be determined based on the user's actual use needs of the electronic device.

[0072] According to the above analysis, the target application of this application can be an AI smart assistant that detects the user's input information to the electronic device, recognizes the user's intention on the input information, and determines the target task corresponding to the user's intention, such as the various voice assistants corresponding to the terminals of major manufacturers. When the user needs to use the electronic device to achieve the target task, the voice signal corresponding to the user's intention can be input, and the voice assistant recognizes the intention of the voice signal, determines the target task, and even determines whether it is necessary to call the target model to respond to the target task, etc. This application does not elaborate on the working principle of the AI ​​smart assistant.

[0073] It should be noted that not all tasks that users request electronic devices to respond to require calling the target model to respond, and the target application may not need to call the target model to respond to the target task, such as starting a control task for a smart home device or an electronic device's music software or other application. The target application determines the control intent of the control task and can directly activate the corresponding software in the electronic device to output an output result that meets the control intent, such as starting the music software to play a certain song; or use the communication link between the electronic device and the smart home device for the control task to send a control instruction to the smart home device to control the smart home device to output an output result that meets the control intent, such as controlling a sweeping robot to work in a user-specified area, starting the air conditioner in a user-specified room to work at a specified temperature, etc.

[0074] Step S12, the target application responds to the target task and outputs an output result for the target task; the output result corresponds to the security level corresponding to the electronic device being in the first operating state, and the security level of the electronic device in the first operating state is lower than the security level of the electronic device in the second operating state.

[0075] According to the above description of the present application, for electronic devices with different operating states, although the security levels corresponding to the first operating state or the second operating state of the electronic device are different, the electronic device can be used to meet the user's task processing needs, and the user does not need to operate the electronic device to enter a specific operating state (such as the second operating state such as the unlocked state) to achieve the target task, thereby enriching the user's control method for using the electronic device to achieve the target task, improving the diversity of task scenarios applicable to the electronic device and the convenience of user input operations.

[0076] However, when the electronic device is in the first operating state corresponding to a lower security level, in order to maximize the security of the output results and reduce the risk of privacy leakage caused by outputting the output results for the target task, the embodiment of the present application proposes that when the electronic device is in different operating states, different output results corresponding to the security level corresponding to the operating state of the electronic device will be output through the target application in response to the same target task, thereby ensuring the security of the output results as much as possible while meeting the user's target task processing needs.

[0077] Among them, for the convenience of description, this application records the security level corresponding to the electronic device being in the first operating state as the first security level, and the security level corresponding to the electronic device being in the second operating state as the second security level, and the second security level is higher than the first security level. As analyzed above, this application can consider that the security of the use environment of the electronic device in the first security level is lower than the security of the use environment of the electronic device in the second security level. Therefore, when the electronic device is in the first operating state to achieve the target task requested by the user, the security requirements for the output result of the output for the target task are higher. When this application responds to the target task and obtains the output result to be output for the target task, it uses a reasonable processing method to make the output result correspond to the first security level, and ensures that the output of the output result meets the security requirements corresponding to the first security level.

[0078] Based on the above analysis, when the electronic device is in the first operating state, although the first security level of the electronic device is lower than the second security level corresponding to the second operating state of the electronic device, in the embodiment of the present application, the target application running on the electronic device can still directly respond to the target task. During the target task response process, the target application can call the target model to respond to the target task, and use the understanding ability and content generation ability of the target model to obtain the content used to determine the output result for the target task, and output the output result for the target task that meets the output security requirements of the electronic device at the first security level, so as to safely and reliably meet the user's target task processing needs. After the electronic device is lost or robbed, the illegal user requests the electronic device to achieve the target task. The electronic device in the first operating state executes the method proposed in the embodiment of the present application, which can ensure as much as possible that the private content is not leaked through the output.

[0079] In one possible implementation, the target application can determine whether it is necessary to call the target model in response to the target task by analyzing the processing intention of the target task, such as combining the implementation method for the target task configured in the electronic device; it can also determine whether it is necessary to call the target model for implementation based on the task type of the target task. For this purpose, the various task types that can call the target model can be pre-configured, and whether the target model needs to be called can be determined by detecting whether the target task belongs to one of the task types, etc. In addition, in the process of obtaining the target task, it is also possible to determine whether it is necessary to call the target model in response to the target task by performing intent recognition on the user input information. The present application does not limit the method for determining whether the target task response process needs to call the target model, and how to call the target model in response to the target task. The calling implementation method does not limit it, and it can be implemented based on the pre-configured control logic of the electronic device.

[0080] To sum up, in the scenario where the user requests the electronic device to automatically achieve the target task, even if the electronic device is in the first operating state, although the current first security level of the electronic device is lower than the second security level corresponding to the second operating state of the electronic device, the user does not need to switch the electronic device to the second operating state. The target application in the running state directly responds to the target task in the first operating state, and outputs the output result for the target task. While meeting the user's target task processing needs and improving the diversity of responding to target tasks, it is ensured that the output result corresponds to the first security level, thereby improving the security of responding to target tasks and reducing the risk of privacy leakage due to the electronic device outputting the output result for the target task at a lower security level.

[0081] Exemplarily, if the first operating state of the electronic device is a locked screen state and the second operating state is an unlocked state, when the electronic device is in the locked screen state, the user inputs task request information for the target task to the electronic device, such as inputting voice signals such as "What are the itineraries for today" or "What new messages are there in the mailbox?", as described above in relation to the task processing method proposed in the embodiment of the present application, the AI ​​smart assistant in the running state (i.e., a target application) performs intent recognition on the voice signal input by the user, and determines that the task to be responded to that meets the user's intent is a target task that needs to be implemented by calling an AI large model (i.e., a target model), without the user unlocking the electronic device. When the electronic device is in the locked screen state, the AI ​​smart assistant can automatically call the AI ​​large model to respond to the target task, such as automatically processing a pre-established itinerary or mailbox content, and outputting output results such as today's itinerary content or the content of the email to be processed, and ensures that the output result content corresponds to the security level of the electronic device corresponding to the locked screen state, thereby reducing the risk of leakage caused by the direct output of privacy content such as personal phone numbers, identity information, internal corporate information, or other sensitive information in the user's important itinerary time and location or emails, thereby improving the security and diversity of task processing.

[0082] Reference Figure 2 , is a flowchart of a task processing method provided in the second embodiment of the present application. In order to further improve the output security of the output result content, this embodiment proposes, on the basis of the task processing method described above, adding a security processing step for the output content before executing the output step, such as Figure 2 As shown, the task processing method proposed in this embodiment may include:

[0083] Step S21, if the electronic device is in the first running state and the target application is in the running state, obtaining a target task; the target application can call a target model to respond to the target task;

[0084] Step S22, the target application responds to the target task, and obtains the processing result of the target model responding to the target task;

[0085] In combination with the relevant descriptions of the target application, target model and target task in the above embodiments, when the target task needs to call the target model to respond to implementation, the target application can call the target model through the corresponding API (Application Programming Interface) so that the target model responds to the target task and obtains the processing result for the target task. This process can be implemented based on the task type of the target task and the task processing logic configured by the target model. This application does not limit the implementation method of the target model responding to the target task.

[0086] In a possible implementation, taking the target model as a generative AI application or an AI large model such as an AIGC model as an example, in order to improve the accuracy and stability of the AI ​​large model and ensure that its processing results meet the user's intention (i.e., the processing intention of responding to the target task), this application can be combined with prompt engineering (also known as contextual information) to guide the behavior of the AI ​​large model by optimizing the prompt words (Prompt) that respond to the target task and generate processing results for the target task.

[0087] Based on this, in the process of the target application calling the target model to respond to the target task, the task prompt information for the target task can be determined based on the user input information, and the task prompt information can be input into the target model to guide the target model to respond to the target task, such as understanding the target task content indicated by the task prompt information, and generating a processing result for the target task according to the processing logic of the target task content indicated by the task prompt information.

[0088] Exemplarily, if the user inputs the voice signal "What new messages are there in the mailbox?", the AI ​​smart assistant can identify the user's intention to summarize the content of the emails to be read, and can call the client of the AI ​​big model to read the content of the emails to be read in the mailbox of the user logged in by the electronic device, and combine the summary processing prompt template with the intention of summarizing the content of the emails to be read to form corresponding email content summary prompt information, which is input into the AI ​​big model. According to the summary processing logic indicated by the summary processing prompt words contained in the summary processing prompt template, the content of the emails to be read is understood and summarized to obtain the summary result of the content of the emails to be read, that is, the processing result of the target model in response to the target task.

[0089] It should be noted that, according to the above description of the target task and the target model, the process of obtaining the processing result of the target model in response to the target task includes but is not limited to the content generation process described above. In a possible implementation, the obtained target task or its task description content or task keywords can be directly input into the target model, and the processing result for the target task can be directly generated by using the powerful processing capability of the target model.

[0090] For example, a user inputs the voice signal "What's the weather in Beijing today?" and obtains a weather query task for querying the weather conditions in Beijing. The user inputs the description of the weather query task into the target model or extracts task keywords such as "today", "Beijing" and "weather" from it. The target model can query today's weather data for Beijing from the Meteorological Bureau database based on this, and determine the queried weather data as the processing result of the weather query task, or input the queried weather data into a sentence template to obtain the weather query result.

[0091] It can be seen that for target tasks of different types or contents, the processing logic of the target model in response to the target task may be different, and the present application does not limit the implementation method of step S22. Among them, the processing logic can be configured or adjusted according to actual needs to adjust the implementation method of the target model in response to the target task and improve the accuracy and reliability of the obtained processing results. In addition, the input content and source of the target model in the target task response process include but are not limited to the content that the electronic device has obtained or can obtain but has not yet obtained, such as the input content stored or input locally by the electronic device, or the input content obtained from other devices (such as databases, servers or other interconnected terminal devices, etc.) to which the electronic device is connected, etc., depending on the circumstances.

[0092] Step S23, processing the processing result based on the first security policy to obtain an output result for the target task; the first security policy is a security policy corresponding to a first security level corresponding to the electronic device being in a first operating state, the semantics represented by the output result is the same as the semantics represented by the processing result, and the security of the output result is higher than the security of the processing result;

[0093] Step S24, outputting the output result.

[0094] Combined with the above description of the process of obtaining the processing results, the processing results obtained by the target model in response to the target task may contain privacy content that needs to be protected. If it is directly output as the output result for the target task, it is easy to cause the leakage of privacy content when the electronic device is lost or stolen because the electronic device in the first operating state can execute the task processing method proposed in the embodiment of the present application. In order to solve this problem of privacy content leakage, the embodiment of the present application proposes to further securely process the processing results of the target model in response to the target task to improve the security of the output results obtained by processing.

[0095] Therefore, the embodiment of the present application can process the obtained processing result based on the first security policy to ensure that the semantics represented by the output result for the target task obtained thereby is the same as the semantics represented by the processing result, to ensure that the output result meets the target task processing requirements requested by the user, and after processing the processing result based on the first security policy, the security of the output result obtained thereby is higher than the security of the processing result. In this way, in the case that the processing result contains privacy content, after processing the processing result based on the first security policy, the output result can no longer contain the privacy content, and the electronic device outputs the output result in the first operating state without causing security issues such as leakage of the privacy content.

[0096] The first security policy is a security policy corresponding to the first security level corresponding to the electronic device being in the first operating state, and can be configured according to the security requirements for the output results of the electronic device indicated by the first security level, so as to ensure that the above processing results are processed based on the first security policy, and the output results obtained correspond to the first security level and meet the security requirements indicated by the first security level. This application does not limit the content of the first security policy.

[0097] In a possible implementation, in a scenario where the determined or called sensitive word library represents various privacy contents that need to be protected, the above-mentioned first security policy may include a filtering policy based on the sensitive word library. In the implementation process of step S23, the obtained processing result may be processed based on the filtering policy of the sensitive word library, that is, the obtained processing result is filtered based on the sensitive word library to filter out sensitive words in the obtained processing result that belong to the sensitive word library to ensure that the obtained output result no longer contains the sensitive word, such as various user identity information and device identification information. Compared with the security of the processing result obtained by the target model in response to the target task, after being processed by the filtering policy, the security of the obtained output result is improved.

[0098] In another possible implementation, the first security strategy may also include an anonymization processing strategy. For the privacy content that needs to be anonymized, such as user information and identity information in the processing results obtained by the target model in response to the target task, information shielding processing can be performed directly, such as filtering, blocking or hiding the user information and identity information (such as replacing them with pre-configured security information or specific symbols (such as xxx or so-and-so, etc.) to obtain a more secure output result, so that when the output result is output, the device information and user information of the electronic device are no longer output, so as to protect privacy content such as device information and user information. Among them, the privacy content that needs to be anonymized can be determined according to relevant protection laws or internal protection regulations of the enterprise, and this application does not limit the content of the anonymization processing strategy and the method for determining the privacy content.

[0099] In another possible implementation, the processing result obtained by the target model in response to the target task can also be processed using a privacy protection strategy based on differential privacy technology, that is, the processing result is processed based on differential privacy technology, and the same type of random noise data is added to the processing result to blur the processing result, so as to achieve the purpose of protecting privacy content. Optionally, each sentence in the processing result can be processed into a word embedding vector, and the sampled Laplace noise is added to each word embedding vector to generate a perturbation word embedding vector. After that, by searching for the nearest neighbor vector of the perturbation word embedding vector (which can be implemented using the Euclidean distance operator), the original sentence is replaced by the sentence corresponding to the nearest neighbor vector, so as to avoid the direct output of the original sensitive sentence and achieve the purpose of privacy protection.

[0100] In the actual application of the present application, the first security strategy may include but is not limited to the three strategies described above. One or more combined strategies can be reasonably selected according to the security requirements for the output results of the target task output by the electronic device in the first operating state as indicated by the first security level to form a first security strategy, so that after obtaining the processing result of the target model in response to the target task, the processing result is first processed based on the security strategy, such as shielding user information such as user name, telephone number, address, social account and / or work unit information in the processing result, as well as privacy information that the company does not disclose to the public, bank transaction flow and other property data, so that the processed processing result no longer contains these shielded privacy contents, and the security of the output result for the target task obtained thereby is higher than the security of the processing result directly obtained by the target model.

[0101] Reference Figure 3 , which is a flow chart of the task processing method provided in the third embodiment of the present application. In order to further improve the output security of the output result content, this embodiment proposes, on the basis of the task processing method described above, adding a security processing step for the target data required to respond to the target task obtained by the target model before the target application responds to the target task step, so as to solve the privacy leakage problem caused by the privacy content introduced by the target model input, so that the output result for the target task output by the electronic device in the first operating state contains relevant privacy content.

[0102] Based on this, Figure 3 As shown, the task processing method proposed in this embodiment may include:

[0103] Step S31, if the electronic device is in the first running state and the target application is in the running state, obtaining a target task; the target application can call a target model to respond to the target task;

[0104] Step S32, the target application obtains target data required to respond to the target task;

[0105] Step S33, processing the target data based on a first security policy to obtain security data; the first security policy is a security policy corresponding to a first security level corresponding to the electronic device being in a first operating state, and the security of the security data is higher than the security of the target data;

[0106] Combined with the above description of the target application calling the target model in response to the target task, in the scenario where the target task is a task of summarizing the content of emails to be read, the target data obtained by the embodiment of the present application may include the content of emails to be read in the user's mailbox (such as the content of the email body and the content of the attachment file, etc.) and email content summary instructions (which can be obtained based on the analysis of the voice signal input by the user), etc. The content of the email to be read may include but is not limited to the privacy content described above. If the target data is directly provided to the target model, so that the target model responds to the target task based on the target data, the output result for the target task is likely to contain privacy content. Once the electronic device is lost or stolen or used illegally, directly outputting an output result such as the email content summary result in this way will cause serious security and privacy leakage problems.

[0107] In order to solve the above-mentioned security and privacy leakage problems, the embodiment of the present application proposes that after obtaining the target data required to respond to the target task, before providing it to the target model, the target data is first processed based on the first security policy to obtain security data with better security than the target data, and then the subsequent call input operation to the target model is performed. Combined with the relevant description of the first security policy in the above embodiment, if the target data contains sensitive words in the sensitive word library (or the target data matches the sensitive word library), the filtering strategy based on the sensitive word library can be executed on the target data to filter and delete each sensitive word in the target data (such as sensitive words in the sensitive word library or sensitive words matching therewith), to ensure that the obtained security data no longer contains sensitive words belonging to the sensitive word library, or the security data does not match the sensitive word library, thereby improving the data security of the input target model.

[0108] If the target data contains private content that needs to be anonymized, such as user information and device information, the user information and device information in the target data can be shielded based on the anonymization strategy to prohibit the obtained security data from containing these private contents. Of course, if the security requirements indicated by the first security level are very high, a privacy protection strategy based on differential privacy technology can be adopted to add random noise data to the obtained target data to blur the content of the target data, thereby protecting the private content in the target data, so that the security data obtained no longer contains the same private content, thereby improving the security of the security data.

[0109] It is understandable that the first security policy corresponding to the first security level used in the embodiment of the present application may include but is not limited to the multiple policies described above, such as first anonymizing the target data, and then filtering the anonymized data based on the sensitive word library, such as removing punctuation marks in the anonymized data, using the DFA (Deterministic Finite Automaton) algorithm to identify and shield sensitive words in the sensitive word library, and then performing differential privacy processing on the sensitive word shielded data to obtain security data, so as to maximize the security of the security data. In order to ensure the accuracy of the output result, during the execution of the first security policy, the policy content can be reasonably selected, such as not executing the privacy protection policy based on the differential privacy technology, ensuring that the semantics represented by the security data is the same as the semantics represented by the target data, thereby ensuring that the target model responds to the target task, and the output result meets the processing requirements of the user's request for the target task.

[0110] It should be noted that the sensitive word library based on each policy included in the first security policy mentioned above, as well as the privacy content that needs to be anonymized, can be customized or adjusted according to the user's own security requirements, and the implementation process is not described in detail in this application. Optionally, for each of the above security policies, the corresponding filter / security check model can be configured to execute the corresponding security policy on the input target data to avoid sending privacy content to the target model, so that the target model generates an output result containing privacy content.

[0111] Step S34, providing the security data to the target model, so that the target model responds to the target task based on the security data and outputs an output result for the target task.

[0112] Regarding the target application calling the target model, transmitting the obtained security data to the target model, and the target model understanding and analyzing the target data and outputting the output results for the target task, the implementation process can refer to the description of the corresponding part of the above embodiment, which will not be described in detail in this embodiment.

[0113] It can be seen that in an embodiment of the present application, when the electronic device is in a first operating state and the target application is in an operating state, if a target task is obtained, even if the current first security level of the electronic device is lower than the second security level corresponding to the electronic device being in a second operating state, the target application can still call the target model to respond to the target task. In this implementation process, after obtaining the target data required to respond to the target task, the present embodiment will process the target data based on the first security policy to obtain security data with higher security than the target data, and provide the security data to the target model to respond to the target task. Compared with the target model directly responding to the target task based on the target data, the present embodiment responds to the target task based on the security data, thereby improving the security of the output result for the target task, and the present application is no longer limited to calling the target model to respond to the target task only when the electronic device is in the second operating state, thereby improving the diversity of task processing and the convenience of users using electronic devices to achieve target tasks.

[0114] In some embodiments, when the target model responds to the target task based on the security data, after obtaining the processing result, the processing result can be processed based on the first security policy to obtain the output result for the target task. Figure 4 As shown in the flow chart, this embodiment can perform security processing on the target data required to respond to the target task, and perform security processing on the processing result obtained by the target model in response to the target task, that is, perform double security processing operations on the input content and generated content of the target model based on the first security policy, and reliably prohibit the output result obtained for the target task from containing privacy content, thereby ensuring the security of the electronic device outputting the output result in the first operating state. This double security processing process can be implemented in combination with the corresponding steps of the above embodiments 2 and 3, which will not be described in detail in this embodiment.

[0115] Reference Figure 5 , is a flowchart of the task processing method provided in the fifth embodiment of the present application. This embodiment can describe an optional detailed implementation process of the target application calling the target model to respond to the target task in the task processing method described in the above embodiment, such as the target model responding to the target task based on the existing knowledge base. The other steps of the task processing method can be described in accordance with the corresponding part of the context embodiment, and this embodiment will not be described in detail here. Based on this, Figure 5 As shown, the task processing method proposed in this embodiment may include:

[0116] Step S51, if the electronic device is in the first running state and the target application is in the running state, obtaining a target task; the target application can call a target model to respond to the target task;

[0117] Step S52, the target application configures the target model based on the first configuration parameter, so that the target model responds to the target task based on the characteristic information of the public knowledge base and generates an output result for the target task; the first configuration parameter is a configuration parameter corresponding to the security level corresponding to the electronic device being in the first operating state, and the security of the output result generated by the target model in response to the target task based on the characteristic information of the public knowledge base is higher than the security of the output result generated by the target model in response to the target task based on the characteristic information of the personal knowledge base;

[0118] Step S53: output the output result for the target task.

[0119] In the embodiments of the present application, a public knowledge base refers to a collection of knowledge shared within an organization / enterprise or society, which is usually maintained and used by multiple people. It can be built and maintained using tools such as knowledge management systems, content management systems, databases, search engines, etc., and can usually include the organization / enterprise's knowledge assets, industry standards, public data sets, encyclopedias, etc. A personal knowledge base can refer to a collection of content such as knowledge, information, experience, and skills accumulated by a user in order to improve work efficiency and promote personal development. It can be built and maintained using tools such as note-taking software, personal databases, and cloud storage services, and contains the user's personal notes, documents, project materials, learning materials, personal experiences, etc.

[0120] It can be seen that the public knowledge base has characteristics such as openness, sharing and standardization, while the personal knowledge base has characteristics such as personalization, privacy and flexibility, which makes the security of the feature information of the public knowledge base itself higher than the security of the feature information of the personal knowledge base itself. Outputting feature information from the public knowledge base usually does not have the problem of privacy content leakage, while outputting feature information from a personal knowledge base may have the risk of privacy leakage.

[0121] Based on this, the embodiment of the present application proposes that when the electronic device is in the first operating state, when the target model responds to the target task, in order to improve the security of the output result, the target model can respond to the target task based on the characteristic information of the public knowledge base. In order to realize this task processing, a first configuration parameter corresponding to the first security level corresponding to the electronic device being in the first operating state can be determined for realizing this task processing method. In this way, when it is necessary to call the target model to respond to the target task, the target model can be configured based on the first configuration parameter first, so that the configured target model responds to the target task based on the characteristic information of the public knowledge base, and generates an output result for the target task. It can be seen that the content constituting the output result comes from the characteristic information of the public knowledge base, or is generated based on the characteristic information of the public knowledge base. Since the characteristic information of the public knowledge base does not belong to private content, the output result is avoided due to the introduction of private content into the knowledge base content, thereby avoiding the privacy leakage problem caused by the introduction of private content into the knowledge base content, thereby improving the security of the output result.

[0122] In some embodiments, reference Figure 6 The flowchart of the sixth embodiment of the task processing method shown in the figure shows that when the electronic device is in the second operating state, the target application calls the target model to respond to the target task. Since the second security level of the electronic device is currently high, it is usually the owner of the electronic device who uses the electronic device to request to achieve the target task. In this case, the target model can be allowed to query the personal knowledge base to respond to the target task and generate an output result for the target task. In order to implement this task processing method, the second configuration parameter corresponding to the second security level corresponding to the second operating state of the electronic device can be determined in advance for implementing the task processing method. The second configuration parameter is different from the first configuration parameter. This application does not limit the content of these two configuration parameters and their determination and implementation methods.

[0123] Afterwards, when the electronic device is in the second operating state, after obtaining the target task, the target model is called to respond to the target task based on the feature information of the personal knowledge base, and an output result for the target task is generated, such as Figure 6 As shown, the output result can be directly output at this time. In order to improve the output security, the target model obtains the processing result in response to the target task based on the characteristic information of the personal knowledge base, processes the processing result based on the second security policy corresponding to the second security level, obtains the output result for the target task, and then outputs the output result. It can be understood that, in combination with the above description of the first security level and the second security level, the security processing strength of the processing result by the first security policy is equal to or higher than the security processing strength of the processing result by the second security policy. This application does not limit the content of the second security policy, such as filtering strategies based on sensitive word libraries.

[0124] In other embodiments, in the process of the target model responding to the target task based on the characteristic information of the knowledge base (such as the personal knowledge base and the public knowledge base mentioned above), the user input information (such as the target data mentioned above) can also be processed based on the security policy of the security level of the corresponding electronic device, and the privacy content contained in the user input information can be identified and shielded to avoid the output result generated in response to the target task from containing the privacy content or other privacy content related to it, thereby improving the security of the output result. Of course, the present application can also process the data at the input and output ends of the target model based on the corresponding security policy according to the method described above, so as to reliably avoid the output result for the target task outputted by the electronic device from involving privacy content, and effectively ensure the security of the output result.

[0125] Reference Figure 7 , which is a flowchart of the task processing method provided in the seventh embodiment of the present application. Based on the task processing method described in any of the above embodiments proposed in the present application, it is also possible to further detect whether the process of the electronic device obtaining the target task is in a trusted environment, determine whether the current electronic device is used by its owner or other authorized legitimate users, avoid using the electronic device to respond to the target task in an untrusted environment, and improve the security of task processing. For other steps of the task processing method, the description of the corresponding embodiment of the above text application can be found, and this embodiment will not be described in detail here.

[0126] Based on this, Figure 7 As shown, the task processing method proposed in this embodiment may include but is not limited to the following steps:

[0127] Step S71, if the electronic device is in the first running state and the target application is in the running state, obtaining a target task; the target application can call a target model to respond to the target task;

[0128] Step S72, obtaining target parameters;

[0129] Step S73, if the target parameter determines that the target condition is satisfied, the target application responds to the target task and outputs an output result for the target task; the target parameter determines that the target condition is satisfied, indicating that the process of obtaining the target task is in a trusted environment.

[0130] In the embodiment of the present application, the target parameter may be a relevant parameter of the electronic device for determining whether the process of obtaining the target task is in a trusted environment, and may be determined based on the judgment logic for determining whether the process of obtaining the target task is in a trusted environment. The target condition may be determined based on the judgment logic for the corresponding target parameter. For different types of target parameters, the judgment logic executed using the target parameter is different, and the corresponding target condition content obtained is different. The present application does not limit the content and acquisition method of each target parameter and its corresponding target condition.

[0131] In actual applications, the target parameters may change as the environment in which the electronic device obtains the target task changes. During the process of the electronic device obtaining the target task, the target parameters remain basically stable. In this way, before responding to the target task, it is possible to first determine whether the corresponding target conditions are met based on the target parameters of the current electronic device. Only when the target parameters are determined to meet the target conditions can the target application running the electronic device respond to the target task, as described in the above embodiments, by calling the target model to respond to the target task, ensuring that the target model responds to the target task and outputting the output result for the target task is performed in a trusted environment, thereby reducing the risk of privacy leakage caused by outputting the output result through the electronic device in the first operating state.

[0132] In one possible implementation, when an electronic device stores location information of one or more trusted environments, such as the address of a trusted environment (i.e., a trusted environment) of the owner of the electronic device, such as a home address or a personal office address, the target parameters obtained by the electronic device may include the location information of the environment in which the electronic device is located, that is, the location information of the environment in which the electronic device is performing a target task. Afterwards, it can be determined whether the location information matches the location information of each trusted environment stored by the electronic device. If the location information of the current electronic device matches the location information of any stored trusted environment, it can be considered that the process of the electronic device performing the target task is in a trusted environment, that is, the target parameters meet the target conditions, so as to continue to execute the subsequent steps of the task processing method proposed in the embodiment of the present application.

[0133] On the contrary, if the obtained location information does not match the location information of each trusted environment stored in the electronic device, it is considered that the process of the electronic device obtaining the target task is in an untrusted environment (a risky environment that is prone to privacy leakage). In order to achieve privacy protection, the electronic device can directly prohibit the target application from responding to the target task, and can also output risk warning information as needed to remind the owner of the electronic device to check the safety of the electronic device. Among them, the acquisition of the above-mentioned location information can be achieved by using the location sensor of the electronic device itself. The implementation process is not described in detail in this embodiment. The present application can also flexibly configure the representation of the location information according to actual conditions, such as latitude and longitude coordinates (if it is the location information of the stored trusted environment, it can be a coordinate range value) or street house number or the layout location of each office within the enterprise, etc., to meet the location matching requirements of various types of trusted environments.

[0134] In some other embodiments, if the electronic device stores the device identification of one or more bound trusted devices, the present application does not limit the implementation method of how the electronic device binds the device identification of the trusted device based on the user identity information, and it can be determined according to the device type of the trusted device and the communication method between the trusted device and the electronic device. For example, a trusted device such as a headset or a bracelet that communicates based on Bluetooth can be implemented through pairing authentication, and the binding between the electronic device and the headset / bracelet can be achieved after the pairing authentication is passed, and the device identification of the headset / bracelet is recorded, such as the device UUID (Universally Unique Identifier) ​​or the unique identification configured in the electronic device for the bound trusted device, such as the device number / name, etc. The present application does not limit the content of the device identification.

[0135] In this case, the target parameters obtained by the electronic device may include the device identification of at least one output device that is currently communicatively connected to the electronic device. For example, during the process of establishing a communication connection between the electronic device and the output device, the device identification of the output device that is successfully communicatively connected to the electronic device can be determined and stored, so that after obtaining the target task, the stored device identification of each output device can be read; or a request can be sent to each output device that is communicatively connected to request the device identification of the corresponding output device, etc. The present application does not limit the implementation method of how to obtain the device identification of each output device that is communicatively connected to the electronic device.

[0136] Afterwards, the device identifications of each output device that is connected to the electronic device can be matched with the device identifications of each stored trusted device. If at least one trusted device identification is matched, it means that the corresponding output device that the electronic device is currently connected to is a trusted device, and it is considered that the process of the electronic device obtaining the target task is in a trusted environment, that is, the target parameter determines that it meets the target condition, and the target application can respond to the target task. On the contrary, if the target parameter determines that it does not meet the target condition, it will be prohibited to respond to the target task, thereby avoiding the risk of private content disclosure due to the loss or theft of the electronic device or malicious abuse.

[0137] It can be seen that the present application can determine whether the process of the electronic device obtaining the target task is in a trusted environment by, but not limited to, the different target parameters described above. As analyzed above, the location information of the trusted environment can be configured to strengthen the presence condition of the owner, i.e., the user, of the electronic device currently being used. The configuration of the location information of the trusted environment can be achieved by converging or even locking the spatial range that allows the target model to respond to the target task input / output privacy content. In order to further determine the target condition of whether the owner of the electronic device is present when the electronic device is used to perform the task processing method, as analyzed above, it can also be determined whether the electronic device is connected to a trusted device. Combining one or two of these implementation methods, it is determined that the electronic device obtains the target task in a trusted environment, and the target application will continue to respond to the target task; otherwise, it is determined that the electronic device is currently in a risky environment, and the process is terminated directly, and it is prohibited to respond to the target task in a risky environment, so as to avoid the risk of privacy leakage caused by outputting the output results for the target task in a risky environment.

[0138] Among them, combined with the relevant description of the target parameters above, if the target parameters are determined not to meet the target conditions, in addition to ending the response to the target task and prohibiting the call of the target model to respond to the processing steps of the target task, a security level switching prompt information can also be output to prompt the user to switch the electronic device to the second operating state, and then respond to the target task after improving the security level of the electronic device, that is, by responding to the operating state switching operation input by the user, the electronic device is controlled to switch from the first operating state to the second operating state, so that the target model continues to respond to the target task, and outputs the output result for the target task corresponding to the security level corresponding to the electronic device being in the second operating state.

[0139] In the task processing method proposed in each embodiment of the above text application, when it is necessary to output the output result for the target task, the obtained output result can be sent to an output device (external device) that is communicatively connected to the electronic device, or an output device that the electronic device itself has, and the output device outputs the output result for the target task. In the process of selecting an output device to output the output result, the output result can be transmitted to an output device that supports an output method that matches the data format according to the data format of the output result, so that the output device can reliably output the output result for the target task. For example, if the output result is a display content composed of text and / or image data, the display screen of the electronic device can be controlled to output the display content; if the output result is audio data, the audio player of the electronic device itself or the Bluetooth headset connected in communication can be controlled to output audio data, etc.

[0140] Optionally, according to the trusted environment determination method described above, it is determined that the electronic device is in a trusted environment of the owner, and the output result for the target task can be output through the output device such as the display screen or audio player or video player of the electronic device itself, or the output device in the trusted environment interconnected by the electronic device. In the case of determining that the electronic device is communicatively connected to at least one trusted device, a trusted device whose output mode matches the data format of the output result can be selected to output the output result, but it is not limited to the output implementation method of the output result described in the embodiment of the present application.

[0141] In the implementation process of the task processing method described in each embodiment above, after obtaining the output result for the target task, before outputting it, the output result can also be subjected to a security check. If the security check result passes, the output result for the target task is output according to but not limited to the output method described above. If the security check result fails, the response to the target task can be ended, the output of the output result can be prohibited, or a security level switching prompt message can be output to prompt the user to increase the security level of the electronic device by switching the operating state of the electronic device. In combination with the target task response implementation method described above, an output result corresponding to the new security level (such as switching to the second security level corresponding to the second operating state) is output. The implementation process is not described in detail in this application.

[0142] In actual applications, since a large target model requires a lot of resources to run, in order to avoid insufficient resources of the electronic device causing the target model to fail to run or run slowly, affecting its processing efficiency in responding to the target task, the target model can be deployed on the server side. In this way, in the process of calling the target model to respond to the target task, the electronic device can send the target task or the target data required for its implementation to the server, and the server inputs the received target data into the target model, feeds back the processing result, and controls the electronic device in the first operating state to output the output result corresponding to its first security level. The task processing process saves the processing resources of the electronic device, ensures the response efficiency to the target task and the output security of the output result, but is not limited to this method of calling the target model to achieve the target task.

[0143] Based on the task processing method described in the above embodiments, as well as the security processing of the input and output ends of the target model, and the related description of the electronic device obtaining the target task in a trusted environment, the following will take the scenario where a user uses a mobile phone in lock screen state and calls the AIGC model to implement the task of summarizing the content of unread emails as an example to illustrate. In order to maximize the security of the email content summary results, the security processing methods described in the above embodiments can be combined to achieve it.

[0144] Based on this, refer to Figure 8The task processing method flow chart shown in the figure shows that when the mobile phone is in the lock screen state, the user inputs the voice signal "What new messages are there in the mailbox?", and the voice assistant in the running state in the electronic device determines the target task requested by the user by identifying the intention of the voice signal, and needs to call the AIGC model to respond to the target task. It can detect whether the electronic device currently has trusted location information, such as determining whether it is in a trusted environment based on the current location information of the electronic device as mentioned above. If not, end the response; if so, request the electronic device to connect to a trusted device to determine whether the electronic device is communicatively connected to at least one trusted device. If not, end the response; if so, the obtained unread email content can be processed based on the first security policy, such as Figure 8 As shown, target data such as unread email content can be input into a local input filter that supports the first security policy. If sensitive content in the target data is successfully filtered and security data is obtained, the security data is provided to the target model, and in response to the email content aggregation task, aggregated email content is obtained.

[0145] Afterwards, the summarized email content can be graded and filtered based on the first security policy. If it is a processing result involving privacy content, the summarized email content can be filtered / shielded and other security processing can be performed on it, and the mobile phone can also output a more secure email content summary result when the screen is locked; if it is a processing result that does not involve privacy content, such as weather, calendar or social news query content, there is no need to perform security processing and the corresponding query content can be directly output.

[0146] It can be seen that in the actual application of the embodiments of the present application, the mobile phone is in the lock screen state, and the user triggers the voice assistant of the mobile phone to call the client of the AI ​​large model. By analyzing the voice signal input by the user, it is determined that the task requested by the user to the electronic device to implement is a target task that needs to be implemented by calling the AI ​​large model. The present application can still respond to the target task and output the request reply content, that is, the output result, corresponding to the security level of the electronic device in the lock screen state. The user does not need to unlock the mobile phone, which improves the user experience. Moreover, since the content of the output result corresponds to the lower security level in the lock screen state, combined with appropriate security strategies, it can effectively reduce the risk of privacy content being leaked.

[0147] A task processing method provided by an embodiment of the present application is introduced above, and a device for executing the above task processing method will be introduced below.

[0148] Reference Fig. 9 , is a schematic diagram of the structure of a task processing device provided in Example 1 of the present application. Fig. 9 As shown, the task processing device may include:

[0149] A target task obtaining module 901 is used to obtain a target task if the electronic device is in a first running state and the target application is in a running state, and the target application can call a target model to respond to the target task;

[0150] An output module 902, configured for the target application to output an output result for the target task in response to the target task;

[0151] The output result corresponds to a security level corresponding to the electronic device being in a first operating state; and the security level of the electronic device being in the first operating state is lower than the security level of the electronic device being in a second operating state.

[0152] In one possible implementation, Fig.10 As shown, the task processing device may further include:

[0153] The processing result obtaining module 903 is used to obtain the processing result of the target model in response to the target task;

[0154] An output result obtaining module 904, configured to process the processing result based on the first security policy to obtain the output result;

[0155] Among them, the first security policy is a security policy corresponding to the security level corresponding to the electronic device being in a first operating state; the semantics represented by the output result is the same as the semantics represented by the processing result, and the security of the output result is higher than the security of the processing result.

[0156] In one possible implementation, Fig.11 As shown, the task processing device may further include:

[0157] The target data acquisition module 905 is used to obtain the target data;

[0158] The security data acquisition module 906 is used to process the target data based on a first security policy to obtain security data; wherein the first security policy is a security policy corresponding to the security level corresponding to the electronic device being in a first operating state, and the security of the security data is higher than the security of the target data.

[0159] The safety data transmission module 907 is used to provide the safety data to the target model so that the target model responds to the target task based on the safety data.

[0160] In one possible implementation, Fig.12 As shown, the task processing device may further include:

[0161] A target parameter obtaining module 908 is used to obtain target parameters;

[0162] A determination module 909 is configured to trigger the output module 902 to execute the step of the target application outputting an output result for the target task in response to the target task if the target parameter determines that the target condition is satisfied;

[0163] The target parameter determination that meets the target condition represents that the process of obtaining the target task is in a trusted environment.

[0164] Optionally, the above-mentioned determination module 909 may include at least one of the following determination units:

[0165] A first determining unit, configured to, if the target parameter includes location information of an environment in which the electronic device is located, determine that the location information matches location information of any trusted environment stored by the electronic device;

[0166] The second determining unit is configured to determine, if the target parameter includes a device identification of at least one output device communicatively connected to the electronic device, that the device identification matches a device identification of a trusted device bound to the electronic device.

[0167] Optionally, the output module 902 may include:

[0168] A configuration unit, configured to configure the target model based on a first configuration parameter so that the target model responds to the target task based on feature information of a public knowledge base;

[0169] An output unit, used to output an output result for the target task;

[0170] Among them, the first configuration parameter is a configuration parameter corresponding to the security level corresponding to the electronic device being in the first operating state; the security of the output result generated by the target model in response to the target task based on the characteristic information of the public knowledge base is higher than the security of the output result generated by the target model in response to the target task based on the characteristic information of the personal knowledge base.

[0171] In the actual application of this application, the first security strategy includes at least one of the following:

[0172] A filtering strategy based on a sensitive word library is used to prohibit the output result from containing sensitive words in the sensitive word library;

[0173] An anonymization processing strategy for shielding user information and device information to prohibit the output result from including device information and user information of the electronic device;

[0174] Privacy protection strategy based on differential privacy technology.

[0175] Optionally, the task processing device may further include:

[0176] A verification module, used for performing security verification on the output result;

[0177] An end module is used to end the response to the target task if the security check result fails, so as to prohibit the output of the output result for the target task, or a prompt information output module is used to output the security level switching prompt information, so as to control the electronic device to switch from the first operating state to the second operating state by responding to the operating state switching operation input by the user, so that the target model responds to the target task and outputs the output result for the target task corresponding to the security level corresponding to the electronic device being in the second operating state.

[0178] An embodiment of the present application also provides a computer program product including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements any task processing method provided in the embodiment of the present application.

[0179] When the above-mentioned computer-readable instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer-readable instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.

[0180] A computer-readable storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any task processing method provided in an embodiment of the present application.

[0181] The computer-readable storage medium may be any available medium that can be stored by a computer or a data storage device such as a training device or a data center that includes one or more available media. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state disk (SSD)).

[0182] Reference Fig.13 , is a schematic diagram of the hardware structure of an electronic device proposed in Example 1 of the present application, such as Fig.13 As shown, the electronic device may include: at least one processor 131 and at least one memory 132, wherein:

[0183] The processor 131 and the memory 132 can communicate via a bus. The bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.13 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0184] The memory 132 can store computer program instructions for implementing the task processing method provided in the embodiments of the present application, as well as program codes for supporting the operation of the target application. The processor 1312 loads the computer program instructions stored in the memory 132 to execute / run, thereby implementing the task processing method provided in the corresponding embodiment. The implementation process can refer to the description of the corresponding part of the method embodiment above.

[0185] In the embodiment of the present application, the processor 131 may include one or more processing units, such as a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), a digital signal processor (DSP), a hardware circuit (such as an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a modem processor, and any one or more of the above processors.

[0186] The memory 132 may be a non-volatile memory, such as a hard disk drive (HDD) or a solid-state drive (SSD), etc., or a volatile memory (volatile memory), such as a random-access memory (RAM). The memory is any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 132 in the embodiment of the present application may also be a circuit or any other device that can implement a storage function, for storing program instructions and / or data.

[0187] In actual applications, in order to achieve communication connection between electronic devices and external devices (such as the above-mentioned output devices or servers, etc.), electronic devices may also include wireless communication elements that support any one or more communication standards / protocols, as well as wired communication elements such as serial / parallel ports, one or more models of USB interfaces, etc.

[0188] It should be understood that Fig.13 The structure of the electronic device shown does not constitute a limitation on the electronic device in the embodiment of the present application. In practical applications, the electronic device may include Fig.13 More components as shown, or combinations of certain components. Fig.14 As shown, the electronic device may also include a display unit (used to display the output results for the target task, various prompt information, etc.), a microphone (used to collect voice signals input by the user), a speaker, various sensors, a power supply module, a radio frequency unit (used to send and receive information, which may include an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (Low Noise Amplifier, LNA) and a duplexer, etc., to achieve data transmission between the electronic device and the base station), various communication elements, external ports and other input units (such as a touch screen, one or more of a mouse key, a function key, a joystick or a trackball, etc.) / output units, etc., which can be determined according to the processing function requirements, and this application does not list them one by one.

[0189] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the device embodiments provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.

[0190] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course can also be implemented by dedicated hardware including dedicated integrated circuits, dedicated CPUs, dedicated memories, dedicated components, etc. In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, it can be implemented in whole or in part in the form of a computer program product.

[0191] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the devices and electronic devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method part.

Claims

1. A task processing method, the method comprising: If the electronic device is in a first running state and the target application is in a running state, obtaining a target task, the target application can call a target model to respond to the target task; The target application outputs an output result for the target task in response to the target task; Wherein, the output result corresponds to the security level corresponding to the electronic device being in the first operating state; The security level of the electronic device in the first operating state is lower than the security level of the electronic device in the second operating state.

2. The method according to claim 1, before outputting the output result for the target task, the method further comprises: Obtaining a processing result of the target model in response to the target task; Process the processing result based on the first security policy to obtain the output result; Wherein, the first security policy is a security policy corresponding to a security level corresponding to the electronic device being in a first operating state; The semantics represented by the output result is the same as the semantics represented by the processing result, and the security of the output result is higher than the security of the processing result.

3. The method according to claim 1, before the target model obtains the target data required to respond to the target task, the method further comprises: obtaining the target data; Process the target data based on the first security policy to obtain security data; wherein, The first security policy is a security policy corresponding to a security level corresponding to the electronic device being in a first operating state, and the security of the security data is higher than the security of the target data; The safety data is provided to the target model so that the target model responds to the target task based on the safety data.

4. The method according to claim 1, wherein the target application responds to the target task by: The target application configures the target model based on the first configuration parameter so that the target model responds to the target task based on the feature information of the public knowledge base; Wherein, the first configuration parameter is a configuration parameter corresponding to the security level corresponding to the electronic device being in the first operating state; The security of the output result generated by the target model in response to the target task based on the feature information of the public knowledge base is higher than the security of the output result generated by the target model in response to the target task based on the feature information of the personal knowledge base.

5. The method according to claim 1, wherein the target application outputs an output result for the target task in response to the target task, and the method further comprises: Get target parameters; If the target parameter determines that the target condition is satisfied, executing the target application in response to the target task to output an output result for the target task; The target parameter determination that meets the target condition represents that the process of obtaining the target task is in a trusted environment.

6. The method according to claim 2 or 3, wherein the first security policy comprises at least one of the following: A filtering strategy based on a sensitive word library is used to prohibit the output result from containing sensitive words in the sensitive word library; An anonymization processing strategy for shielding user information and device information to prohibit the output result from including device information and user information of the electronic device; Privacy protection strategy based on differential privacy technology.

7. The method according to claim 5, wherein the target parameter determines that the target condition is satisfied, comprising at least one of the following: If the target parameter includes location information of the environment in which the electronic device is located, determining that the location information matches location information of any trusted environment stored by the electronic device; If the target parameter includes a device identification of at least one output device communicatively connected to the electronic device, it is determined that the device identification matches a device identification of a trusted device bound to the electronic device.

8. The method according to claim 1, before outputting the output result for the target task, the method further comprises: Performing security verification on the output result; If the security check result is failed, the response to the target task is terminated to prohibit the output of the output result for the target task, or a security level switching prompt message is output to control the electronic device to switch from the first operating state to the second operating state in response to the operating state switching operation input by the user, so that the target model responds to the target task and outputs the output result for the target task corresponding to the security level corresponding to the electronic device being in the second operating state.

9. A task processing device, comprising: A target task acquisition module, used for acquiring a target task if the electronic device is in a first running state and the target application is in a running state, wherein the target application can call a target model to respond to the target task; An output module, configured for the target application to output an output result for the target task in response to the target task; Wherein, the output result corresponds to the security level corresponding to the electronic device being in the first operating state; The security level of the electronic device in the first operating state is lower than the security level of the electronic device in the second operating state.

10. An electronic device, comprising: at least one memory and at least one processor, wherein: A memory for storing computer program instructions and program codes for supporting the operation of a target application; The processor is used to load and execute the computer program instructions to implement the following steps: If the electronic device is in a first running state and the target application is in a running state, a target task is obtained, and the target application can call a target model to respond to the target task; The target application outputs an output result for the target task in response to the target task; Wherein, the output result corresponds to the security level corresponding to the electronic device being in the first operating state; The security level of the electronic device in the first operating state is lower than the security level of the electronic device in the second operating state.

Images