Data processing method and device, equipment and storage medium

Abstract

The invention provides a data processing method and device, equipment and a storage medium. First task information can be obtained. Then, first prompt information can be generated based on the first task information and sent to the language model. After the first operation information sent by the language model is received, the first operation indicated by the first operation information can be executed, the execution state information of the first operation information is stored, and the execution state information corresponding to the first operation information is sent to the client. Thus, the user triggers further feedback information on the client for the execution state information, the execution state information can be called back, the second prompt information is generated according to the execution state information and the feedback information, and the second prompt information is sent to the language model. Therefore, after the feedback information is received, new prompt information can be generated according to the stored execution state information and the feedback information, so that the language model analyzes the feedback information on the basis of the execution state information.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a data processing method, device, equipment and storage medium. Background Art

[0002] Executing corresponding actions based on user instructions is a common application scenario. In particular, with the development of computer technology, it is possible to respond to user instructions through models. Specifically, users can issue instructions multiple times. The model can obtain the user's instructions and perform corresponding actions based on the user's instructions.

[0003] In some possible implementations, the model may determine a task chain based on the user's instructions. The task chain includes multiple operations that have a sequence. Moreover, the execution of some operations in the task chain may depend on related operations triggered by the user. Summary of the invention

[0004] The present application provides a data processing method, apparatus, device and storage medium.

[0005] In a first aspect, the present application provides a data processing method, characterized in that the method comprises:

[0006] Get the first task information;

[0007] Sending first prompt information to a language model, wherein the first prompt information is generated based on the first task information;

[0008] Receiving first operation information returned by the language model;

[0009] Execute a first operation indicated by the first operation information;

[0010] storing execution status information of the first operation information;

[0011] Sending execution status information of the first operation information to the client;

[0012] In response to receiving feedback information regarding the execution status information, calling back the execution status information;

[0013] generating second prompt information to be sent to the language model based on the execution state information and the feedback information, and receiving second operation information returned by the language model;

[0014] The first operation information and the second operation information are used to indicate operations having a sequential execution order in the task chain corresponding to the first task information.

[0015] In some possible implementations, the method further includes:

[0016] Receiving second task information sent by the client;

[0017] In response to the second task information being unassociated with the execution status information, deleting the execution status information.

[0018] In some possible implementations, the method further includes:

[0019] Sending third prompt information to the language model, where the second prompt information is generated based on the third task information.

[0020] In some possible implementations, the method further includes:

[0021] Execute the second operation indicated by the second operation information;

[0022] The execution status information is updated based on the second operation information and the execution result of the second operation.

[0023] In some possible implementations, determining the execution state information corresponding to the first operation information includes:

[0024] A first content card is generated according to the first operation information, the first content card includes a target operation control, the target operation control is used to trigger a target operation, and the feedback information is generated based on the target operation.

[0025] In some possible implementations, obtaining the first task information includes:

[0026] The first task information is obtained through interaction information between the user and the intelligent assistant.

[0027] In some possible implementations, the interaction information is an instant messaging session message;

[0028] The method further comprises:

[0029] The execution status information is displayed through an instant messaging session corresponding to the intelligent assistant.

[0030] In a second aspect, the present application provides a data processing device, characterized in that the device comprises:

[0031] An acquiring unit, used for acquiring first task information;

[0032] A first sending unit, configured to send first prompt information to the language model, so that the language model returns first operation information, wherein the first prompt information is generated based on the first task information;

[0033] A receiving unit, configured to receive first operation information returned by the speech model; a processing unit, configured to execute a first operation indicated by the first operation information;

[0034] a storage unit, configured to store execution status information of the first operation information;

[0035] A second sending unit, configured to send execution status information of the first operation information to the client;

[0036] The processing unit is further configured to, in response to receiving feedback information regarding the execution status information, call back the execution status information;

[0037] The first sending unit is further used to generate second prompt information to be sent to the language model based on the execution state information and the feedback information, so that the language model returns second operation information;

[0038] The first operation information and the second operation information are used to indicate operations having a sequential execution order in the task chain corresponding to the first task information.

[0039] In some possible implementations, the acquisition unit is further used to receive third task information sent by the client; the processing unit is further used to delete the execution status information in response to the third task information being irrelevant to the execution status information.

[0040] In some possible implementations, the first sending unit is further configured to send second prompt information to the language model, where the second prompt information includes the second task information.

[0041] In some possible implementations, the processing unit is further configured to execute a second operation indicated by the second operation information; and update the execution status information based on the second operation information and an execution result of the second operation.

[0042] In some possible implementations, the processing unit is specifically used to generate a first content card according to the first operation information, the first content card includes a target operation control, the target operation control is used to trigger a target operation, and the feedback information is generated based on the target operation.

[0043] In some possible implementations, the acquisition unit is specifically configured to acquire the first task information through interaction information between the user and the intelligent assistant.

[0044] In some possible implementations, the interaction information is an instant messaging session message; and the processing unit is specifically configured to display the execution status information through the instant messaging session corresponding to the intelligent assistant.

[0045] In a third aspect, the present application provides an electronic device, including:

[0046] one or more processors;

[0047] a storage device having one or more programs stored thereon,

[0048] When the one or more programs are executed by the one or more processors, the one or more processors implement any method described in the first aspect.

[0049] In a fourth aspect, the present application provides a computer-readable medium having a computer program stored thereon, wherein when the program is executed by a processor, any method described in the first aspect is implemented.

[0050] In a fifth aspect, the present application provides a computer program product, which, when executed on a device, enables the device to execute the method described in the first aspect.

[0051] The present application provides a data processing method, apparatus, device and storage medium. Specifically, when executing the data processing method, first task information can be obtained. The first task information corresponds to a task chain, and the task chain corresponding to the first task information includes at least two operations with a sequential order. Then, first prompt information can be generated based on the first task information and sent to a language model so that the language model returns first operation information according to the first prompt information. After receiving the first operation information sent by the language model, the first operation indicated by the first operation information can be executed, the execution status information of the first operation information can be stored, and the execution status information corresponding to the first operation information can be sent to the client. In this way, the user triggers further feedback information on the execution status information on the client, and the execution status information can be called back, and the second prompt information can be generated according to the execution status information and the feedback information, and the second prompt information can be sent to the language model. The language model can return the second operation information according to the second prompt information. Among them, the first operation information and the second operation information are respectively used to indicate two operations with a sequential execution order in the task chain corresponding to the first task information. That is, if an operation in the task chain depends on the user's feedback information for a certain operation, then after obtaining the operation information of the operation, the stored execution status information can be called back. In this way, after receiving the feedback information, new prompt information can be generated based on the stored execution status information and feedback information, so that the language model can parse the feedback information based on the execution status information. In this way, storing the execution status information allows the user to trigger other operations unrelated to the execution status information, and can also resume the execution of the task chain after obtaining feedback information related to the execution status information. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 A flowchart of a data processing method provided in an embodiment of the present application;

[0054] Figure 2 A schematic diagram of the structure of a data processing device provided in an embodiment of the present application;

[0055] Figure 3 A schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0056] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be construed as being limited to the embodiments described herein. Instead, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.

[0057] It should be understood that the various steps described in the method implementation of the present application can be performed in different orders and / or performed in parallel. In addition, the method implementation may include additional steps and / or omit the steps shown. The scope of the present application is not limited in this respect.

[0058] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0059] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different objects, devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0060] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0061] With the development of computer technology, some models are endowed with the ability to parse instructions issued by users. Accordingly, in some application scenarios, this ability can be used to improve user experience. Specifically, users can input instructions through voice or text on the client. The model can obtain the user's instructions through the network and determine the user's intention by parsing the user's instructions. In the above process, the information issued by the user can be called instruction information, and the information obtained by the model parsing the instruction information can be called reply information.

[0062] In some possible implementations, the reply information can be used to indicate the execution of an operation. That is, after obtaining the reply information, one or more tools can be called according to the reply information to perform corresponding operations according to the user's intention. Therefore, in some possible implementations, the instruction information sent by the user can be called task information, and the reply information fed back by the model can be called operation information.

[0063] In order to further improve the user experience, some models also have the ability to learn context. For example, models with natural language processing (NLP) capabilities, such as language models (LM), can combine historical task information and / or historical operation information to determine the corresponding results.

[0064] In some possible implementations, a language model can be used to parse the task information sent by the user, and perform corresponding operations according to the user's instructions. Specifically, the user can input task information in the form of natural language text through the client. The task information is forwarded to the language model. The language model can parse the task information and determine the user's intention. The language model can feedback a natural language description indicating the execution steps of the task information (for example, the operation information referred to below) so as to perform corresponding operations according to these execution steps. In addition, since the model has the ability to learn context, if the user sends other task information before obtaining new instructions, the model can learn historical task information and historical operations in order to process the newly received task information.

[0065] In some possible implementations, the instruction information given by the user may be relatively complex. The user's instruction information may correspond to multiple operations. In an embodiment of the present application, the multiple operations corresponding to the instruction information may be referred to as a task chain. The task chain may include multiple operations that are executed in sequence. In order to complete the user's intention, each operation in the task chain may be executed in sequence.

[0066] For ease of introduction, for an operation in a task chain, an operation that is executed before the operation in the task chain can be called a preceding operation, and an operation that is executed after the operation can be called a succeeding operation.

[0067] Optionally, the determination and execution of some operations in the task chain may depend on the user's feedback on other operations. That is, the determination and execution of the N+1th operation (N is a positive integer) in the task chain may depend on the user's feedback on the execution result of the Nth operation in the task chain. Therefore, when executing operations according to the task chain, after executing the Nth operation, you can wait for the user's feedback on the Nth operation. Before obtaining the user's feedback, the N+1th operation may not be executed. After obtaining the user's feedback, you can continue to execute the N+1th operation.

[0068] For example, assume that the task chain includes operation A and operation B. The language model can determine the operation information A2 based on the task information A1 sent by the user. The execution subject (for example, the client or server that receives the user instruction information) can execute operation A based on the operation information A2, and feedback the execution result to the user through the client. Then, the user can further trigger task information B1 based on the execution result of operation A. After obtaining the task information B1, the language model can combine the task information A1, the operation information A2 and the task information B1 for context learning, so as to determine the operation information B2 based on the task information B1. Similarly, the execution subject can execute operation B based on the operation information B2.

[0069] In the above process, the operation information B2 corresponding to operation B is determined by the language model based on the operation information A2 and the task information B1. Therefore, in order to ensure the normal execution of the tasks in the task chain, the task information A1 and the operation information A2 can be stored. In this way, after the task information B1 is obtained, the operation information B2 can be determined based on the stored task information A1 and operation information A2.

[0070] However, if the user enters other task information other than task information B1 after entering task information A1, the already stored task information A1 and operation information A2 may be deleted. The task chain cannot continue to execute. Therefore, in order to maintain the continuity of the task chain, the user's operation can be restricted. For example, it can be displayed that the user can only trigger feedback related to the previous operation in the task chain. That is, in the above example, after executing operation A, the user can be restricted to only trigger subsequent operations on the execution result of operation A. However, this restricts the user's operation and has a greater impact on the user experience.

[0071] In order to solve the problems of the prior art, an embodiment of the present application provides a data processing method, which is described in detail below in conjunction with the drawings in the specification.

[0072] See also Figure 1 , Figure 1 An embodiment of the present application provides a flow chart of a data processing method, and the embodiment of the present application can be applied to an application scenario in which an action is to be performed according to a user's instruction. Among them, the data processing method can be executed by a data processing device. The data processing device can be implemented by software. Optionally, the device may include a model for determining a natural language description of the execution steps of the indicated task information according to user instructions, or the device may be associated with a model for determining a natural language description of the execution steps of the indicated task information according to user instructions. Optionally, the above-mentioned model may include, for example, an LM model. In an embodiment of the present application, the model may be referred to as a language model.

[0073] Optionally, the data processing device can be integrated into the server. The server can be a server of instant messaging (IM) software, which is used to provide corresponding services for the client of IM software. The client can be a software client or a web client. The user can send messages to the server by using the client. The server can receive messages sent by the client and perform data processing. The server can also send data processing results to the client so that the client can display the data processing results to the user. Optionally, the data processing device can be integrated with a module in the server for implementing the action execution function. Optionally, the server corresponding to the data processing device can include a language model, or the server corresponding to the data processing device can exchange information with a device for implementing the language model.

[0074] For ease of introduction, the following description is made by taking the case where the data processing device is integrated into the server as an example.

[0075] like Figure 1 As shown, the method specifically comprises the following steps:

[0076] S101: Obtain first task information.

[0077] When a user wants to call certain functions through a language model, the user can send first task information through a client. Accordingly, the data processing device can obtain the first task information sent by the user. The first task information indicates the user's intention to control the software to perform one or more tasks. By parsing the first task information, one or more operations can be performed according to the user's intention.

[0078] In an embodiment of the present application, the first task information is used to indicate multiple tasks. That is, the first task information corresponds to a task chain. The task chain corresponding to the first task information can be called a first task chain, and the first task chain includes multiple operations with a sequential execution order. It should be noted that the operations in the first task chain are performed by the data processing device, which is different from the user-triggered operations such as the first operation described below. Optionally, in order to distinguish between user-triggered operations and operations in the first task chain that need to be performed by the data processing device, the user-triggered operation can be referred to as an action.

[0079] It should be noted that the operations in the task chain may be triggered by the user's instructions. For example, the user can input instructions multiple times, and each time the user's output instructions are obtained, the corresponding operations can be performed according to the user's instructions. These operations constitute a task chain. Alternatively, the operations in the task chain need to be gradually obtained by multiple reasonings of the language model. In other words, when the user issues the first instruction, the task chain may not be determined. With the instructions issued by the user or with the multiple calls of the language model, the operations in the task chain are gradually determined. That is, when the first task information is obtained, the data processing device cannot determine the first task chain. After the data processing device performs the first operation and the second operation according to the user's instructions (such as the first task information and feedback information), the first operation and the second operation can be collectively referred to as the first task chain.

[0080] Optionally, the first task information may be sent by the client to the data processing device via a network. The user may trigger an operation on the client to control the client to send the first task information to the server. The client of the software may determine the first task information based on an action triggered by the user. In an embodiment of the present application, the action triggered by the user and corresponding to the first task information may be referred to as the first action.

[0081] The first action is introduced below.

[0082] In some possible implementations, the first action may include an action of a user sending a text message. The first task information may be obtained based on text information input by the user. The user may give instructions to the data processing device by inputting text information. For example, the user may send the first task information to the server by sending a message to the intelligent assistant. The intelligent assistant may be a pre-set interface of a service associated with a language model. Optionally, if the data processing device runs on the server of the IM software, the intelligent assistant may be a pre-set IM dialogue robot.

[0083] In an embodiment of the present application, the first task information may be sent via an IM message. The IM message sent by the user to the intelligent assistant may include text information, and the text information indicates the first task information. Accordingly, the action of the user sending the IM message is the first action, and the first task information may include the text in the IM message.

[0084] In some possible implementations, the intelligent assistant service may be provided to the user in the form of a fixed IM session. The fixed IM session may be referred to as a target IM session, which is used to receive IM messages sent by the user. That is, the user may initiate a target IM session and send IM messages through the target IM session. The data processing device may obtain the first task information according to the target IM message sent by the user to execute the first task chain.

[0085] Optionally, the target IM session may be initiated by the user. For example, the user may search for the smart assistant through the contact list or service number list in the IM client to initiate the target IM session. Specifically, after the user turns on the corresponding function, the IM contact corresponding to the smart assistant may be added to the user's contact list, and the user may initiate the target IM session through the IM contact. Alternatively, after the user turns on the corresponding function, the target IM session may be added to the user's IM session list. Accordingly, the user may find the target IM session by browsing or searching to send the target IM message.

[0086] Optionally, the above-mentioned user input text information may include text information in text format, or may include text information obtained by converting audio input by the user. Specifically, the user may input text information through a text box, for example, through an IM message input box corresponding to the IM software. Alternatively, the user may also input voice information and send it to the data processing device in the form of a voice IM message. The data processing device may parse the voice IM message and convert the voice input by the user into text information.

[0087] In the implementation described above, the first action includes an action of information input by the user. In some other possible implementations, the first action may also include an action of the user triggering a control to generate and send the first task information to the server. Specifically, one or more controls may be displayed to the user, and different controls may be used to trigger different actions, corresponding to different task information. Optionally, the correspondence between the controls and the task information may be pre-set in the data processing device. Accordingly, the user may trigger a single-click operation (or other operation) on the control corresponding to the first task information, and the data processing device may determine the first task information based on the operation triggered by the user and the correspondence between the control and the first task information.

[0088] S102: Sending first prompt information to the language model.

[0089] After acquiring the first task information, the data processing device may generate first prompt information according to the first task information, and send the first task information to the language model. The first prompt information may be prompt information. Optionally, if the first task information is not the first information sent to the language model in the first task chain, the first prompt information may also include historical conversation information.

[0090] In the embodiment of the present application, the first prompt information is generated based on the first task information. Optionally, the first prompt information may include the first task information.

[0091] Optionally, if the data processing apparatus and the language model do not run on the same computing device or computing device cluster, the data processing apparatus may send the first task information to the language model through an interface provided externally by the language model.

[0092] S103: Receive first operation information returned by the language model.

[0093] After obtaining the first task information, the language model can infer the first task information to determine the first operation information corresponding to the first task information, and send the first operation information to the data processing device. The first operation information can correspond to an operation in the first task chain. For ease of explanation, the operation corresponding to the first operation information is referred to as the first operation below.

[0094] For example, user A can send an IM message to the target IM dialogue corresponding to the intelligent robot through the IM client. The content of the IM message includes "Please help me schedule a meeting at 10 o'clock tomorrow morning, including user B and user C". The IM message is sent to the IM server through the network, and then prompt information is sent to the language model based on the content of the message.

[0095] The language model can infer based on the prompt information and determine that the corresponding task chain includes the action "create a meeting schedule" and the action "notify user B and user C". Accordingly, the first operation may include the operation "create a meeting schedule", and the first operation information may include the operation information corresponding to the action "create a meeting schedule".

[0096] S104: Execute the first operation indicated by the first operation information.

[0097] After acquiring the first operation information, the data processing device may perform the first operation according to the instruction of the first operation information. By performing the first operation, the execution status information of the first operation information may be obtained. Optionally, the execution status information of the first operation information may include the execution result of the first operation.

[0098] In some possible implementations, the data processing device may perform the first operation by calling a tool. The tool may be a pre-configured software tool for implementing a specific data processing function. Optionally, multiple tools may be pre-set. After obtaining the first operation information, the corresponding tool may be called according to the first operation information to perform the first operation.

[0099] In some possible implementations, the first operation may include processing data and displaying the processed data to the user. Optionally, the data processing device may first call a data processing tool to process the data, and then call an information sending tool to display the processed data to the user.

[0100] That is, after acquiring the first operation information, the data processing device may first determine a virtual tool for processing the data according to the first operation information, and then process the data by calling the virtual tool.

[0101] For example, assuming that the first operation information indicates that the user's intention is to display pre-stored information A, after obtaining the first operation information, the data processing device can call the information search tool according to the first operation information to search for the pre-stored information A from the pre-stored data.

[0102] For another example, assume that the first operation information indicates that the user's intention includes searching for information related to question B. Then, after obtaining the first operation information, the data processing device may call a search engine-like tool according to the first operation information to search for information related to question B. Then, the data processing device may send the information related to question B to the user.

[0103] After the data processing is completed, the data processing device can also call a tool for sending messages (such as IM messages) to send the processed data to the client. Accordingly, the processed data is the execution status information of the first operation information.

[0104] The above example is used for illustration. Assume that the first task information sent by the user includes an IM message "Please help me schedule a meeting at 10 a.m. tomorrow, including user B and user C". Then after obtaining the first operation information, the data processing device can first call a virtual tool of the schedule type to create a meeting schedule at 10 a.m. tomorrow, and set the participants of the meeting to "user A, user B and user C".

[0105] S105: Storing the execution status information of the first operation information.

[0106] After executing the first operation according to the instruction of the first operation information, the data processing device can store the execution status information of the first operation information. The execution status information of the first operation information is used to describe the first operation. According to the execution status information of the first operation information, the post-operation of the first operation in the first task chain can be continued to be executed.

[0107] Optionally, the data processing device may store the execution status information in a memory of the data processing device, or in other storage devices of the data processing device. After obtaining feedback information for the execution status information, the client may call back the execution status information.

[0108] S106: Sending execution status information of the first operation information to the client.

[0109] After the first operation is performed according to the instruction of the first operation information, the data processing device may send the execution status information of the first operation information to the client, so that the user using the client can understand the execution status information of the first operation.

[0110] In some possible implementations, the execution status information may be displayed in the form of a content card. The content card that displays the execution status information may be referred to as a first content card. The first content card corresponds to a portion of the display area of ​​the IM client. The background color of the first content card may be different from other display areas, and / or the first content card may be separated from other display areas by a border. The first content card includes an information display area and a first operation control. The information display area may be used to display information to the user, for example, it may be used to display the above-mentioned processed data. The first operation control may be used to trigger a second operation.

[0111] Optionally, multiple types of content cards may be pre-set. After the execution state information is determined according to the first operation information, a content card corresponding to the execution state information may be selected to display the execution state information. For example, if the execution state information is schedule information, a content card for displaying the schedule information may be selected, and the specific content in the execution state information may be filled into the content card to obtain a first content card.

[0112] According to the foregoing introduction, the first task information can be obtained based on an IM message sent by a user. Accordingly, the execution status information of the first operation information can be sent to the user based on the IM message. Specifically, the first task information sent by the user and the execution status information can be received and sent to the user through the same IM session. Accordingly, the client can display the processed data to the user through the target IM session. For example, an IM message is generated based on the execution status information, and the IM message may include an identifier of the target IM session or an identifier of the intelligent assistant. After receiving the IM message, the client can determine the target IM session based on the identifier of the target IM session or the identifier of the intelligent assistant, and display the execution status information to the user in the target IM session.

[0113] In an embodiment of the present application, the first task chain may include multiple operations such as a first operation and a second operation. The second operation is a post-operation of the first operation. In addition, the second operation is performed based on feedback information of the execution status information. In other words, after the execution status information is displayed to the user, the user may trigger further operations with respect to the execution status information. According to the operation triggered by the user with respect to the execution status information, the client may send feedback information with respect to the execution status information to the server.

[0114] For ease of introduction, the operation triggered by the user for the execution status information is referred to as the second action below. That is, the user can first perform the first action so that the target model infers the first operation information and causes the data processing device to perform the first operation. Then, the user can perform the second action for the execution status information corresponding to the first operation to trigger the target model to obtain the second operation information, and the data processing device performs the second operation according to the second operation information. In this process, the first operation and the second operation constitute the first task chain. For an introduction to the second action, please refer to the following.

[0115] In some possible implementations, the second action may include an action of a user triggering a control. Accordingly, when the execution status information is displayed, an operation control corresponding to the second action may be displayed. The operation control may be referred to as a target operation control. The target operation control is used to trigger a target operation. The second operation performed by the user may include triggering an operation through a target operation control. According to the target operation triggered by the user, the data processing device may obtain feedback information. Optionally, if the execution status information includes a first content card, the target operation control may be a control on the first content card.

[0116] S107: In response to receiving feedback information regarding the execution status information, calling back the execution status information.

[0117] After the client obtains the execution status information, the client can display the execution status information to the user. Based on the execution status information, the user can trigger further operations. Based on the operations triggered by the user, the data processing device can obtain feedback information for the execution status information. The feedback information represents the user's feedback on the execution status information, for example, it can reflect the user's opinion on the execution status information. After obtaining the feedback information, the data processing device can call back the execution status information.

[0118] The following first introduces the feedback information.

[0119] As described above, the feedback information is feedback on the execution status information. Accordingly, the second action may be a user-triggered action related to the execution status information. For example, the feedback information may be used to confirm the execution status information, and the second action may be an action for confirming the execution status information.

[0120] In some possible implementations, the second action may include an action of triggering a target operation control. Alternatively, in some other possible implementations, the second action may also include an action of inputting instruction information. For example, the second action may include an action of a user sending an IM message. Specifically, if the second action is an action for confirming the execution status information, then the second action may specifically include an action of sending an IM message "confirmation content" to the intelligent assistant. The data processing device may determine feedback information for the execution status information using the IM message "confirmation content".

[0121] Optionally, in some scenarios, the user may not be satisfied with the execution status information and needs to adjust the execution status information or reject the execution status information. In these scenarios, the above feedback information can also be used to reject the confirmation of the execution status information. Accordingly, the first content card can also include a rejection operation control, which is used to trigger a rejection operation, and the rejection operation is used to reject the confirmation of the execution status information.

[0122] After obtaining the feedback information, the data processing device may call back the stored execution status information to generate the second prompt information and send it to the language model.

[0123] S108: Generate second prompt information to be sent to the language model based on the execution state information and the feedback information.

[0124] After calling back the execution status information, the data processing device may generate second prompt information according to the execution status information and the feedback information, and send the second prompt information to the language model. Optionally, the second prompt information may include the execution status information and the feedback information.

[0125] The language model can parse the feedback information based on the execution state information according to the second prompt information to obtain the second operation information. Then, the language model can return the second operation information to the data processing device. Similar to the first prompt information, the second prompt information can also be prompt information.

[0126] That is to say, the second prompt information sent by the data processing device to the language model includes not only the feedback information, but also the execution status information associated with the feedback information. Therefore, based on the second prompt information, the language model can determine which information the user has provided feedback on, and thus process the feedback information based on the execution status information. Accordingly, since the data processing device can carry the execution status information in the second prompt information, the language model may not store the first operation information. Furthermore, since the data processing device can store the execution status information, even if the user performs other operations on the client, the execution of the first task chain can be continued based on the execution status information and the feedback information.

[0127] After obtaining the feedback information and execution status information sent by the data processing device, the language model can perform context learning based on the execution status information, and determine the natural language description of the execution steps of the indicated task information in combination with the feedback information, so as to reply according to the user's intention. Specifically, the language model can determine the second operation information based on the feedback information, and feed back the second operation information to the data processing device. The second operation information is obtained based on the analysis of the feedback information, and reflects the user's intention when giving the feedback information. Optionally, the second operation information may correspond to the second operation in the first task chain. That is, the data processing device may first execute the second operation in the first task chain. After obtaining the feedback information triggered by the user, the data processing device may continue to execute the second operation in the first task chain according to the analysis instructions of the language model.

[0128] The above example is used for explanation. After obtaining the IM message sent by user A, "Please help me schedule a meeting at 10 am tomorrow, which includes user B and user C", the language model can infer the first task information and determine the corresponding first operation information. According to the first operation information, the data processing device can execute the action of creating a meeting schedule and display the schedule card corresponding to the meeting schedule to the user through the target IM message. In addition, the data processing device can also store the schedule card corresponding to the meeting schedule.

[0129] Assuming that the meeting schedule requires user confirmation, the meeting schedule information cannot be sent to user B and user C respectively before user A confirms the meeting schedule information. In other words, before obtaining the confirmation operation triggered by the user, the operation in the first task chain "sending the meeting schedule information to user B and user C respectively" cannot continue to be executed.

[0130] By viewing the schedule card, the user can determine whether the created meeting schedule meets expectations. If the meeting schedule meets expectations, the user can trigger the confirmation control on the schedule card. Accordingly, the data processing device can obtain feedback information corresponding to the confirmation operation.

[0131] After obtaining the feedback information, the data processing device can call back the stored execution status information and send the execution status information and feedback information to the language model through prompt information. The language model can perform context learning on the execution status information, and after context learning, it can determine the user's opinion on the meeting schedule in combination with the feedback information and continue with subsequent processing.

[0132] The present application provides a data processing method, device, system and storage medium. Specifically, when executing the data processing method, first task information can be obtained. The first task information corresponds to a task chain, and the task chain corresponding to the first task information includes at least two operations with a sequential order. Then, first prompt information can be generated based on the first task information and sent to a language model so that the language model returns first operation information according to the first prompt information. After obtaining the first operation information, the first operation indicated by the first operation information can be executed, the execution status information of the first operation information can be stored, and the execution status information corresponding to the first operation information can be sent to the client. In this way, the user triggers further feedback information on the execution status information on the client, and the execution status information can be called back, and the second prompt information can be generated according to the execution status information and the feedback information, and the second prompt information can be sent to the language model. The language model can return the second operation information according to the second prompt information. Among them, the first operation information and the second operation information are respectively used to indicate two operations with a sequential execution order in the task chain corresponding to the first task information. That is, if an operation in the task chain depends on the user's feedback information for a certain operation, then after obtaining the operation information of the operation, the execution status information of the operation information can be stored. In this way, after receiving the feedback information, new prompt information can be generated based on the stored execution status information and feedback information, so that the language model can parse the feedback information based on the execution status information. In this way, storing the execution status information allows the user to trigger other operations that are not related to the execution status information, and can also resume the execution of the task chain after obtaining feedback information related to the execution status information. In this way, the effect of interrupting the operation can be achieved without restricting the user's operation.

[0133] In the implementation manner described above, the data processing device may store the execution status information so as to call back the execution status information when receiving feedback information regarding the execution status information.

[0134] However, in actual scenarios, after the user views the execution status information, the user may want to perform other operations instead of continuing to execute subsequent operations in the first task chain. To this end, in some possible implementations, after sending the execution status information to the client, the data processing device can also receive second task information sent by the client. The task chain (or task) corresponding to the second task information is different from the task chain (i.e., the first task chain) corresponding to the first task information. In other words, even if the client displays the execution status information to the user, the user can control the data processing device to perform tasks other than the first task chain by triggering the second task information.

[0135] In the implementation described above, the execution status information is stored in the data processing device. That is, every time a user triggers the execution of a task chain through task information, the execution result of a certain action in the task chain may be stored in the data processing device. This will put a lot of pressure on the storage device of the data processing device. In particular, in practical applications, in order to facilitate sending to the language model, the execution results may be stored in the memory of the data processing device. In this way, multiple execution results will occupy more memory of the data processing device, thereby affecting the normal operation of the data processing device.

[0136] Therefore, in order to avoid the above problem, in some possible implementations, the stored execution results may be deleted when task information irrelevant to the execution status information is obtained.

[0137] Specifically, if the user sends the second task information to the data processing device through the client, the data processing device may determine whether the second task information is associated with the execution state information.

[0138] If the second task information is associated with the execution status information, indicating that the user wants to perform an operation related to the first operation, the second task information can be regarded as the above feedback information, the execution status information is called back and the second prompt information is generated.

[0139] If the second task information is not associated with the execution status information, it means that the user wants to perform an operation that is unrelated to the first operation. In this way, the stored execution status information can be deleted. Optionally, third prompt information can be generated based on the second task information, and the third prompt information is sent to the language model so that the language model can parse the user's intention according to the third prompt information.

[0140] The above example is used for explanation. If the user sends an IM message "Please schedule a meeting for me at 10 am tomorrow, which includes user B and user C" through the target IM session, the user can view the schedule card corresponding to the meeting in the IM session. If the user does not trigger the confirmation operation control on the schedule card, but sends other IM messages through the target IM session. And the new IM message sent by the user is not related to the meeting corresponding to the schedule card. Then, the data processing device can delete the relevant information of the stored schedule card.

[0141] Based on a data processing method provided by the above method embodiment, an embodiment of the present application also provides a data processing device, which will be described below in conjunction with the accompanying drawings.

[0142] See also Figure 2 As shown in FIG. 1 , this figure is a schematic diagram of the structure of a data processing device provided in an embodiment of the present application. Figure 2 As shown, the data processing device 200 includes:

[0143] An acquisition unit 210, configured to acquire first task information;

[0144] A first sending unit 220 is used to send first prompt information to the language model, wherein the first prompt information is generated based on the first task information;

[0145] A receiving unit 230, configured to receive first operation information returned by the language model;

[0146] A processing unit 240 is configured to execute a first operation indicated by the first operation information;

[0147] A storage unit 250, configured to store execution status information of the first operation information;

[0148] The second sending unit 260 is used to send the execution status information of the first operation information to the client;

[0149] The processing unit 240 is further configured to call back the execution status information in response to receiving feedback information regarding the execution status information;

[0150] The first sending unit 220 is further configured to generate second prompt information to be sent to the language model based on the execution state information and the feedback information, so that the language model returns second operation information;

[0151] The first operation information and the second operation information are used to indicate operations having a sequential execution order in the task chain corresponding to the first task information.

[0152] In some possible implementations, the acquisition unit 210 is further configured to receive second task information sent by the client; and the processing unit is further configured to delete the execution status information in response to the second task information being unrelated to the execution status information.

[0153] In some possible implementations, the first sending unit 220 is further configured to send third prompt information to the language model, where the third prompt information includes the second task information.

[0154] In some possible implementations, the processing unit 240 is further configured to execute a second operation indicated by the second operation information; and update the execution status information based on the second operation information and an execution result of the second operation.

[0155] In some possible implementations, the processing unit 240 is specifically configured to generate a first content card according to the first operation information, wherein the first content card includes a target operation control, and the target operation control is used to trigger a target operation, and the feedback information is generated based on the target operation.

[0156] In some possible implementations, the acquisition unit 210 is specifically configured to acquire the first task information through interaction information between the user and the intelligent assistant.

[0157] In some possible implementations, the interaction information is an instant messaging session message; the processing unit 240 is specifically configured to display the execution status information through the instant messaging session corresponding to the intelligent assistant.

[0158] Based on a data processing method provided by the above method embodiment, the present application also provides an electronic device, including: one or more processors; a storage device, on which one or more programs are stored, when the one or more programs are executed by the one or more processors, the one or more processors implement the data processing method as described in any of the above embodiments.

[0159] Reference below Figure 3, which shows a schematic diagram of the structure of an electronic device 300 suitable for implementing the embodiment of the present application. The terminal device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (portable android devices), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs (televisions), desktop computers, etc. Figure 3 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0160] like Figure 3 As shown, the electronic device 300 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0161] Typically, the following devices may be connected to the I / O interface 305: input devices 308 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 308 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 309. The communication devices 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 3 The electronic device 3400 having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0162] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 309, or installed from the storage device 308, or installed from the ROM 302. When the computer program is executed by the processing device 301, the above-mentioned functions defined in the method of the embodiment of the present application are executed.

[0163] The electronic device provided in the embodiment of the present application and the data processing method provided in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0164] Based on a data processing method provided in the above method embodiment, an embodiment of the present application provides a computer storage medium on which a computer program is stored, wherein when the program is executed by a processor, the data processing method as described in any of the above embodiments is implemented.

[0165] It should be noted that the computer-readable medium mentioned above in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0166] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (Hyper Text Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0167] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0168] The computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device executes the data processing method.

[0169] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0170] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0171] The units involved in the embodiments described in this application may be implemented by software or hardware. The name of a unit / module does not limit the unit itself in some cases. For example, a voice data acquisition module may also be described as a "data acquisition module".

[0172] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0173] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0174] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the system or device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0175] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0176] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0177] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A data processing method, characterized in that: The method comprises: Get the first task information; Sending first prompt information to a language model, wherein the first prompt information is generated based on the first task information; receiving first operation information returned by the language model; and performing a first operation indicated by the first operation information; storing execution status information of the first operation information; Sending execution status information of the first operation information to the client; In response to receiving feedback information regarding the execution status information, calling back the execution status information; generating second prompt information to be sent to the language model based on the execution state information and the feedback information, and receiving second operation information returned by the language model; The first operation information and the second operation information are used to indicate operations having a sequential execution order in the task chain corresponding to the first task information.

2. The method according to claim 1, characterized in that: The method further comprises: Receiving second task information sent by the client; In response to the second task information being unassociated with the execution status information, deleting the execution status information.

3. The method according to claim 2, characterized in that The method further comprises: Sending third prompt information to the language model, where the third prompt information is generated based on the second task information.

4. The method according to claim 1, characterized in that: The method further comprises: Execute the second operation indicated by the second operation information; The execution status information is updated based on the second operation information and the execution result of the second operation.

5. The method according to claim 3, characterized in that: The determining of the execution status information corresponding to the first operation information includes: A first content card is generated according to the first operation information, the first content card includes a target operation control, the target operation control is used to trigger a target operation, and the feedback information is generated based on the target operation.

6. The method according to claim 1, characterized in that The obtaining of the first task information includes: The first task information is obtained through interaction information between the user and the intelligent assistant.

7. The method according to claim 6, characterized in that The interactive information is an instant messaging session message; The method further comprises: The execution status information is displayed through an instant messaging session corresponding to the intelligent assistant.

8. A data processing device, characterized in that: The device comprises: An acquiring unit, used for acquiring first task information; A first sending unit, configured to send first prompt information to the language model, so that the language model returns first operation information, wherein the first prompt information is generated based on the first task information; A receiving unit, configured to receive first operation information returned by the language model; a processing unit, configured to execute a first operation indicated by the first operation information; a storage unit, configured to store execution status information of the first operation information; A second sending unit, configured to send execution status information of the first operation information to the client; The processing unit is further configured to, in response to receiving feedback information regarding the execution status information, call back the execution status information; The first sending unit is further used to generate second prompt information to be sent to the language model based on the execution state information and the feedback information, so that the language model returns second operation information; The first operation information and the second operation information are used to indicate operations having a sequential execution order in the task chain corresponding to the first task information.

9. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer-readable medium, characterized in that A computer program is stored thereon, wherein when the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images