Information processing method and apparatus, and device and storage medium

Abstract

Provided in the embodiments of the present disclosure are an information processing method and apparatus, and a device and a storage medium. The method comprises: during the process for performing interaction with a first model regarding a target user input, generating the current instruction for the first model, wherein the current instruction indicates a task to be processed by the first model; determining corresponding degrees of association between a plurality of candidate interaction records and the current instruction, wherein the plurality of candidate interaction records at least comprise the target user input, one or more historical outputs from the first model during interaction, or one or more historical instructions for the first model during interaction; on the basis of the degrees of association of the plurality of candidate interaction records, removing at least one candidate interaction record from the plurality of candidate interaction records, so as to obtain at least one target interaction record; and on the basis of the current instruction and the at least one target interaction record, generating the current input for the first model, so as to process the target user input. In this way, the amount of data involved during interaction can be dynamically controlled, thereby avoiding the situation in which the amount of data is excessively great.

Description

Information processing methods, apparatus, equipment and storage media

[0001] This application claims priority to Chinese Patent Application No. 202410508623.3, filed on April 25, 2024, entitled "Information Processing Method, Apparatus, Device and Storage Medium", the entire contents of which are incorporated herein by reference. Technical Field

[0002] The exemplary embodiments disclosed herein generally relate to the field of computers, and particularly to information processing methods, apparatus, devices, and computer-readable storage media. Background Technology

[0003] With the development of information technology, various terminal devices can provide people with a variety of services in work and life. Applications providing these services can be deployed on these terminal devices. The terminal devices present relevant content and interact with users through the application's user interface to meet various user needs. In some cases, users may initiate information processing requests within the application. Therefore, improving the stability of information processing is a key concern. Summary of the Invention

[0004] In a first aspect of this disclosure, an information processing method is provided. The method includes: during an interaction between a target user input and a first model, generating a current instruction for the first model, the current instruction indicating a task to be processed by the first model; determining the correlation between a plurality of candidate interaction records and the current instruction, the plurality of candidate interaction records including at least the target user input, one or more historical outputs of the first model during the interaction, or one or more historical instructions for the first model during the interaction; removing at least one candidate interaction record from the plurality of candidate interaction records based on the correlation between the candidate interaction records to obtain at least one target interaction record; and generating a current input for the first model based on the current instruction and the at least one target interaction record to process the target user input.

[0005] In a second aspect of this disclosure, an apparatus for information processing is provided. The apparatus includes: an instruction generation module configured to generate a current instruction for the first model during an interaction between a target user input and the first model, the current instruction indicating a task to be processed by the first model; a correlation determination module configured to determine the corresponding correlation between a plurality of candidate interaction records and the current instruction, the plurality of candidate interaction records including at least the target user input, one or more historical outputs of the first model during the interaction, or one or more historical instructions for the first model during the interaction; a record removal module configured to remove at least one candidate interaction record from the plurality of candidate interaction records based on the corresponding correlation of the candidate interaction records to obtain at least one target interaction record; and an input generation module configured to generate a current input for the first model based on the current instruction and at least one target interaction record to process the target user input.

[0006] In a third aspect of this disclosure, an electronic device is provided. The device includes at least one processing unit; and at least one memory coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit. When executed by the at least one processing unit, the instructions cause the electronic device to perform the method of the first aspect.

[0007] In a fourth aspect of this disclosure, a computer-readable storage medium is provided. A computer program is stored on the medium, which, when executed by a processor, implements the method of the first aspect.

[0008] According to a fifth aspect of this disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, it implements the method according to a first aspect of this disclosure.

[0009] It should be understood that the description in this section is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0010] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0011] Figure 1 shows a schematic diagram of an example environment in which embodiments of the present disclosure can be implemented;

[0012] Figure 2 shows a flowchart of an information processing procedure according to some embodiments of the present disclosure;

[0013] Figure 3 illustrates a schematic diagram of an example of selecting historical user input as context for the current user input according to some embodiments of the present disclosure;

[0014] Figure 4 shows a schematic structural block diagram of an apparatus for information processing according to some embodiments of the present disclosure;

[0015] Figure 5 shows a block diagram of an electronic device that can implement one or more embodiments of the present disclosure. Detailed Implementation

[0016] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0017] In the description of embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions may also be included below.

[0018] In this document, unless explicitly stated otherwise, performing a step in response to A does not mean that the step is performed immediately after A, but may include one or more intermediate steps.

[0019] It is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition, use, storage or deletion of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0020] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, relevant users should be informed of the type, scope of use, and usage scenarios of the information involved in this disclosure through appropriate means in accordance with relevant laws and regulations, and authorization from relevant users should be obtained. Among them, relevant users may include any type of rights holder, such as individuals, enterprises, and groups.

[0021] For example, in response to receiving an active request from a user, a prompt message is sent to the relevant user to clearly inform the user that the requested operation will require obtaining and using the user's information, thereby enabling the relevant user to choose whether to provide information to the software or hardware such as the electronic device, application, server, or storage medium that performs the operation of the technical solution disclosed herein based on the prompt message.

[0022] As an optional but non-restrictive implementation, in response to a user's active request, a prompt message can be sent to the user, such as a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide information to the electronic device.

[0023] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure. The activation of digital assistant-related functions, the acquisition of data, the processing and storage of data, etc., in the embodiments of this disclosure shall all require prior authorization from the user and other rights holders associated with the user, and shall comply with the agreements and rules between relevant laws and regulations and rights holders.

[0024] As used in this paper, the term "model" refers to a model that learns the relationship between inputs and outputs from training data, enabling it to generate corresponding outputs for a given input after training. Model generation can be based on machine learning techniques. Deep learning is a machine learning algorithm that processes inputs and provides corresponding outputs using multiple layers of processing units. A neural network model is an example of a deep learning-based model. In this paper, "model" may also be referred to as a "machine learning model," "learning model," "machine learning network," or "learning network," and these terms are used interchangeably.

[0025] Figure 1 illustrates a schematic diagram of an example environment 100 in which embodiments of the present disclosure can be implemented. Environment 100 relates to an application management platform 110, which can support application creation and / or application execution. In some embodiments, the portion of the application management platform 110 used to support application creation may also be referred to as an application creation portion. In some embodiments, the portion of the application management platform 110 used to support application execution may also be referred to as an application execution portion.

[0026] As shown in the figure, the application creation section provides an environment for user 105 to create and publish applications. User 105 can be referred to as the application creation user or creator. In some embodiments, the application creation section can be a low-code platform that provides a collection of tools for application creation. The application creation section can support visual development of various types of applications, allowing developers to skip the manual coding process and accelerate the application development cycle and reduce costs. The application creation section can support any suitable platform for users to develop one or more types of applications, such as an application platform as a service (aPaaS) based platform. Such a platform enables users to efficiently develop applications, enabling operations such as application creation and application function adjustment.

[0027] The application creation component can be deployed locally on user 105's terminal device and / or supported by a server-side device. For example, user 105's terminal device can run a client with the application creation component, which can support interaction between the user and the application creation component provided by the server. When the application creation component runs locally on the user's terminal device, user 105 can directly interact with the local application creation component using the terminal device. When the application creation component runs on a server-side device, the server-side device can provide services to the client running on the terminal device based on the communication connection with the terminal device. The application creation component can present a corresponding page 130 to user 105 based on user 105's actions, to output and / or receive application creation-related information from user 105.

[0028] In some embodiments, the application creation section may be associated with a corresponding database, which stores the data or information required for the application creation process supported by the application creation section. For example, the database may store the code and description information corresponding to the various functional modules that make up the application. The application creation section can also perform operations such as calling, adding, deleting, and updating the functional modules in the database. The database may also store operations that can be performed on different functional blocks. For example, in a scenario where an application needs to be created, the application creation section can call the corresponding functional blocks from the database to build the application.

[0029] In embodiments of this disclosure, user 105 can create and publish target application 120 as needed in the application creation section. Target application 120 can be published to any suitable application runtime section, as long as the application runtime section can support the operation of target application 120. After publication, target application 120 can be operated by one or more end users 145. End user 145 can operate target application 120 through an associated terminal device 146 and thereby interact with application management platform 110. End user 145 can be referred to as the end user of target application 120. In some embodiments, target application 120 may include or be implemented as digital assistant 122.

[0030] Digital assistant 122 can be configured to have intelligent conversational capabilities. In the example shown, digital assistant 122 can be integrated into target application 120, serving as part of target application 120 to assist in task processing within target application 120. In other examples, digital assistant 122 can be configured as a standalone application, such as a web application or other type of application. In such examples, digital assistant 122 and target application 120 can be considered as the same application. Digital assistant 122 is provided to assist users with various task processing needs in different applications and scenarios. During interaction with digital assistant 122, the user inputs interactive messages, and digital assistant 122 responds to the user's input by providing reply messages. Typically, digital assistant 122 can support users inputting questions in natural language and performs tasks and provides replies based on its understanding of natural language input and logical reasoning capabilities.

[0031] In some embodiments, the digital assistant 122 can interact with the end user 145 as a contact. For example, the digital assistant 122 can be implemented in an instant messaging (IM) application. The digital assistant 122 can interact with the end user 145 in a one-on-one chat session. In some embodiments, the digital assistant 122 can interact with multiple users in a group chat session that includes multiple users.

[0032] For each end user 145, the client of the application runtime portion can present an interaction window 142 of the target application 120 or digital assistant 122 in the client interface, such as a conversation window with the digital assistant 122. The end user 145 can enter conversation messages in the conversation window, and the target application 120 can determine the response message from the digital assistant 122 based on the created configuration information and present it to the user in the interaction window 142. In some embodiments, depending on the configuration of the target application 120, the interaction messages with the target application 120 can include multimodal messages, such as text messages (e.g., natural language text), voice messages, image messages, video messages, and so on.

[0033] Similar to the application creation component, the application runtime component can be deployed locally on each end user's (145's) terminal device and / or supported by a server device. For example, the end user's (145's) terminal device can run a client with the application runtime component, which can support interaction between the user and the application runtime component provided by the server. When the application runtime component runs locally on the user's terminal device, the end user (145) can directly interact with the local application runtime component using the terminal device. When the application runtime component runs on a server device, the server device can provide services to the client running on the terminal device based on the communication connection with the terminal device. The application runtime component can present corresponding application pages to the end user (145) based on the user's (145's) actions, outputting and / or receiving application-related information from the user (145).

[0034] In some embodiments, the implementation of at least some functions of the target application 120, and / or the implementation of at least some functions of the digital assistant 122 within the target application 120, may be based on models. During the creation or operation of the target application 120, one or more models 155 may be invoked, such as the capabilities of model 155. In the target application 120, the digital assistant 122 may utilize model 155 to understand user input and provide responses to the user based on the output of model 155.

[0035] During the creation process, the application management platform 110 needs to use model 155 to test the target application 120 to determine whether the running results of the target application 120 meet expectations. During operation, in response to different operation requests from users of the target application 120, the application operation part may need to use model 155 to determine the response results to users.

[0036] Although shown as independent of the application management platform 110, one or more models 155 may run on the application management platform 110 or other remote servers. In some embodiments, model 155 may be a machine learning model, a deep learning model, a learning model, a neural network, etc. In some embodiments, the model may be based on a language model (LM). A language model, by learning from a large corpus, is capable of question answering. Model 155 may also be based on other suitable models.

[0037] The application management platform 110 can run on suitable electronic devices. These electronic devices can be any type of computing-capable device, including terminal devices or server devices. Terminal devices can be any type of mobile terminal, fixed terminal, or portable terminal, including mobile phones, desktop computers, laptop computers, notebook computers, netbook computers, tablet computers, media computers, multimedia tablets, personal communication system (PCS) devices, personal navigation devices, personal digital assistants (PDAs), audio / video players, digital cameras / camcorders, positioning devices, television receivers, radio receivers, e-book devices, gaming devices, or any combination of the foregoing, including accessories and peripherals of these devices or any combination thereof. Server devices can include, for example, computing systems / servers, such as mainframes, edge computing nodes, computing devices in cloud environments, and so on. In some embodiments, the management platform 110 can be implemented based on cloud services.

[0038] It should be understood that the structure and function of environment 100 are described for illustrative purposes only and do not imply any limitation on the scope of this disclosure. For example, although a single user interacting with the application creation section and a single user interacting with the application running section are illustrated, in reality multiple users can access application management platform 110 to each create a digital assistant, and each digital assistant can be used to interact with multiple users.

[0039] Currently, in target applications, due to increased user interactions, rising application complexity, and increasingly sophisticated interactions, the contextual information used by models to process information is also growing. Excessive contextual information may exceed the model's processing capacity; for example, the number of tokens input to the model may exceed its processing limit. This can lead to information loss or processing errors. For instance, it might result in incorrect token usage or the loss of earlier dialogue rounds. A traditional approach is to prune the context, discarding earlier context. However, this can lead to the loss of crucial information from earlier times.

[0040] This disclosure provides an improved information processing scheme. In this scheme, during the interaction between a target user input and a first model, a current instruction is generated for the first model, indicating the task to be processed by the first model. The correlation between multiple candidate interaction records and the current instruction is determined. These candidate interaction records include at least the target user input, one or more historical outputs of the first model during the interaction, and one or more historical instructions for the first model during the interaction. Based on the correlation between the multiple candidate interaction records, at least one candidate interaction record is removed from the multiple candidate interaction records to obtain at least one target interaction record. Based on the current instruction and the at least one target interaction record, a current input is generated for the first model to process the target user input.

[0041] This method allows for the determination of the relevance between multiple candidate interaction records and the current command during the interaction process, and the target interaction record is selected from these candidates based on this relevance. Thus, contextual information is filtered based on relevance during the interaction. This dynamically controls the amount of data during the interaction process, preventing excessive data volume. Furthermore, because relevance is considered in the context filtering, it helps improve the stability of information processing while ensuring accuracy and efficiency.

[0042] Some exemplary embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0043] The task management process described in the embodiments of this disclosure can be implemented on an application management platform, a terminal device with the application management platform installed, and / or a server corresponding to the application management platform. In the examples below, for the sake of discussion, the description is from the perspective of the application management platform, such as the application management platform 110 shown in FIG1. ​​The user interface presented by the application management platform 110 can be presented via the terminal device of user 145, and the application management platform 110 can receive user input via the terminal device of user 145. In some embodiments of this disclosure, user 145 is the terminal user of the target application 120. It should be understood that the user interface presented by the application management platform 110 can also be presented via the terminal device of user 105, and the application management platform 110 can also receive user input via the terminal device of user 105. In some embodiments of this disclosure, user 105 is the creator, manager, or maintainer of the target application 120.

[0044] Figure 2 shows a flowchart of an information processing procedure 200 according to some embodiments of the present disclosure. Procedure 200 can be implemented in the application management platform 110, for example, by the application runtime portion of the application management platform 110. The task processing procedure shown in Figure 2 will be described below with reference to Figure 1.

[0045] In box 210, during the interaction between the target user input and the first model, the application management platform 110 generates a current instruction for the first model, which instructs the first model to handle the task.

[0046] The target user input can be user input from a target user (which can be any suitable user), for example, it could be user input from user 145. The target user input can be any suitable type of user input, such as text, voice, gesture, etc. The application management platform 110 can receive the target user input in any suitable manner, for example, it can receive text-type target user input via an input box, audio-type target user input via a microphone, etc. The target user input can be presented, for example, in an interactive window (e.g., interactive window 142).

[0047] In some embodiments, target user input can instruct operations associated with a structured object. A structured object can be any suitable type of object capable of structurally storing or representing information, including but not limited to data tables, databases, APIs, etc. For example, target user input can instruct operations associated with a data table. Operations associated with a data table can include, for example, adding the corresponding field value of one or more fields, reading the corresponding field value of one or more fields, modifying the corresponding field value of one or more fields, and / or deleting the corresponding field value of one or more fields. That is, target user input can instruct the application management platform 110 to perform CRUD operations on fields in the target data table. Compared to other types of structured objects, interactions with a model targeting a data table typically require providing information about the data table, resulting in a relatively large amount of data provided to the model. Therefore, filtering interaction records is particularly important in data table scenarios.

[0048] The first model can be deployed on the application management platform 110 or on a remote device. The first model can be based on any suitable model architecture, including but not limited to Transformer models, Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Deep Neural Networks (DNNs), and any suitable model. In some embodiments, the first model can be a language model.

[0049] The interaction process between the target user input and the first model is to utilize the first model to execute the task indicated by the target user input, thereby providing a response to the target user input. This interaction process may include multiple rounds of interaction with the first model. For each round of interaction, the application management platform 110 can generate an instruction for the first model (which may be, for example, at least a part of a prompt). For example, if the target user input is "statistically analyze XX data", the application management platform 110 may generate a first instruction in the first round of interaction instructing the first model to retrieve a data table A containing XX data, a second instruction in the second round of interaction instructing the first model to retrieve XX data from data table A, and a third instruction in the third round of interaction instructing the first model to perform statistics on the retrieved XX data. If the current interaction is the third round, then the third instruction is the current instruction for the first model. In each round of interaction, the first model can also generate a corresponding output based on at least the corresponding instruction (i.e., multiple rounds of interaction can correspond to multiple outputs). The application management platform 110 can instruct the first model to complete the task indicated by the target user input through multiple rounds of interaction.

[0050] In box 220, the application management platform 110 determines the correlation between multiple candidate interaction records and the current instruction. The multiple candidate interaction records include at least the target user input, one or more historical outputs of the first model during the interaction process, or one or more historical instructions for the first model during the interaction process. The historical outputs of the first model may be the outputs of the first model in previous rounds of interaction performed in response to the target user input. The historical instructions may be instructions generated by the application management platform 110 in previous rounds of interaction performed in response to the target user input. For example, if the current execution is the third round of interaction, one or more historical outputs may include the outputs of the first model in the first and second rounds of interaction, and one or more historical instructions may include instructions for the first model generated by the application management platform 110 in the first and second rounds of interaction.

[0051] In some embodiments, the application management platform 110 can also utilize a second model to determine the topic of the target user input and the corresponding topics of one or more historical user inputs prior to the target user input. This second model can be the same as the first model or a different model. The second model can also be based on any suitable model structure, including but not limited to Transformer models, CNNs, RNNs, DNNs, etc. Such a second model can also perform operations guided by prompts. The second model can be deployed on the application management platform 110 or on a remote device. The second model can, for example, be a language model. The application management platform 110 can, for example, provide the target user input and one or more historical user inputs to the second model respectively, and obtain the model output of the second model. The model output of the second model can, for example, indicate the topic of the corresponding user input (target user input or historical user input).

[0052] Application management platform 110 can also determine the matching degree between one or more historical user input topics and the target user input topics. Application management platform 110 can determine the matching degree between one or more historical user input topics and the target user input topics using any suitable method. For example, application management platform 110 can determine the matching degree between one or more historical user input topics and the target user input topics based on predetermined rules or algorithms. As another example, application management platform 110 can also utilize machine learning models (e.g., a first model or a second model) to determine the matching degree between one or more historical user input topics and the target user input topics. In some embodiments, application management platform 110 can determine the matching degree between one or more historical user input topics and the target user input topics based on the semantic similarity between the semantics corresponding to the respective topics of one or more historical user input topics and the target user input topics.

[0053] The application management platform 110 can, for example, identify historical user inputs with a matching degree greater than a threshold matching degree as candidate interaction records among multiple candidate interaction records. Referring to FIG3, FIG3 shows a schematic diagram of example 300 according to some embodiments of the present disclosure. As shown in FIG3, if user input N in interaction 310-N corresponds to the current target user input, and user input 1 in interaction 310-1, user input 2 in interaction 310-2, etc. are multiple historical user inputs, the application management platform 110 can determine the topic of the user input in each interaction (e.g., topic 1 corresponding to user input 1 in interaction 310-1, topic 2 corresponding to user input 2 in interaction 310-2, etc., the topics corresponding to multiple historical interaction inputs, and topic N corresponding to user input N in interaction 310-N). The application management platform 110 can determine (320) the matching degree between the topics corresponding to multiple historical user inputs and the topics corresponding to the current target user input. If the matching degree between topic 1 and topic N is greater than the threshold matching degree, and the matching degree between topic 2 and topic N is less than the threshold matching degree, then the application management platform 110 can determine the historical user input (i.e., user input 1) corresponding to subject 1 as a candidate interaction record.

[0054] Therefore, multiple candidate interaction records may include target user input, one or more historical outputs of the first model during the interaction process, one or more historical instructions for the first model during the interaction process, and historical user inputs with a matching degree greater than the threshold matching degree. The one or more historical outputs of the first model during the interaction process are also one or more intermediate interaction results prior to the final execution result mentioned above. The one or more historical instructions are also one or more instructions prior to the current instruction provided to the first model by the application management platform 110.

[0055] Regarding the specific method for determining the correlation between multiple candidate interaction records and the current instruction, in some embodiments, the application management platform 110 can determine the correlation between a given interaction record and the current instruction in at least one of the following dimensions: time dimension, interaction source dimension, or content dimension, for a given interaction record among multiple candidate interaction records, and determine the correlation between the given interaction record and the current instruction based on the correlation in at least one dimension.

[0056] In some embodiments, the first candidate interaction record is generated earlier than the second candidate interaction record, and the correlation between the first candidate interaction record and the second candidate interaction record in the time dimension is lower. That is, the earlier the interaction record is generated, the lower its correlation with the current instruction in the time dimension.

[0057] In some embodiments, the relevance of the target user input under the interaction source dimension is greater than the relevance of one or more historical outputs under the interaction source. Alternatively or additionally, in some embodiments, the relevance of one or more historical outputs under the interaction source is greater than the relevance of one or more historical instructions under the interaction source. That is, the target user input from the target user has the highest relevance under the interaction source dimension, the historical outputs of the first model have a moderate relevance under the interaction source dimension, and the historical instructions for the first model have the lowest relevance under the interaction source dimension. It can be understood that if historical user input is included, the relevance of historical user input under the interaction source can be less than that of the target user input but greater than that of historical outputs.

[0058] In some embodiments, the relevance of a given interaction record in the content dimension is determined based on the matching degree between the topic of the given interaction record and the topic of the current instruction, which can be regarded as the contextual relevance between the given interaction record and the current instruction. The application management platform 110 can also determine the matching degree between the topic of the given interaction record and the topic of the current instruction based on any appropriate method. For example, the application management platform 110 can utilize a second model to determine the topic of the given interaction record and the topic of the current instruction. For example, the application management platform 110 can determine the matching degree between the topic of the given interaction record and the topic of the current instruction based on the semantic similarity between the topics of the given interaction record and the topic of the current instruction. It can be understood that the greater the semantic similarity, the higher the matching degree.

[0059] In some embodiments, if at least one dimension includes multiple dimensions, the application management platform 110 can calculate the correlation between a given interaction record and the current instruction based on the corresponding weights of the multiple dimensions and the corresponding correlation of the given interaction record under the multiple dimensions. Taking a dimension including time, interaction source, and content as an example, if the weight of the time dimension is A, the weight of the interaction source dimension is B, and the weight of the content dimension is C, and the correlation of the given interaction record under the three dimensions of time, interaction source, and content are a, b, and c respectively, then the correlation between the given interaction record and the current instruction is Aa + Bb + Cc. In some embodiments, the weight of the time dimension can be the largest.

[0060] In box 230, the application management platform 110 removes at least one candidate interaction record from the multiple candidate interaction records based on their respective relevance, to obtain at least one target interaction record. For example, the application management platform 110 may remove at least one candidate interaction record with the lowest relevance to obtain at least one target interaction record.

[0061] In some embodiments, the application management platform 110 may, in response to generating a current instruction, directly determine the correlation between multiple candidate interaction records and the current instruction, and remove at least one candidate interaction record from the multiple candidate interaction records based on the correlation between the multiple candidate interaction records. In some embodiments, the application management platform 110 may also determine whether the multiple candidate interaction records will cause the amount of data input to the model to exceed a threshold amount of data, and if it is determined that the multiple candidate interaction records will cause the amount of data input to the model to exceed the threshold amount of data, and the amount of data input at present is lower than the threshold amount of data, remove at least one candidate interaction record from the multiple candidate interaction records based on the correlation between the multiple candidate interaction records.

[0062] Application management platform 110 may, for example, determine whether multiple candidate interaction records will cause the amount of data input to the model to exceed a threshold amount of data before determining the correlation. Application management platform 110 may also determine whether multiple candidate interaction records will cause the amount of data input to the model to exceed the threshold amount of data when or after determining the correlation. This disclosure does not limit the specific timing of determining whether multiple candidate interaction records will cause the amount of data input to the model to exceed the threshold amount of data.

[0063] In box 240, application management platform 110 generates current input for the first model based on the current instruction and at least one target interaction record to process target user input.

[0064] In some embodiments, for each round of interaction, the application management platform 110 can generate the current input for the first model by combining the current instruction corresponding to the current round of interaction and at least one target interaction record. For example, if the current round is the third round, the application management platform 110 can generate the current input (also referred to as the third input) for the user's first model by combining the third instruction corresponding to the current round and at least one target interaction record. The first model can process the current input to generate the corresponding output.

[0065] In some embodiments, the application management platform 110 can also provide the user with the processing results of the first model for the target user's input via an interactive window. The application management platform 110 may not provide the user with multiple intermediate interaction results corresponding to multiple interaction rounds (the output of the first model corresponding to the last round or the last few rounds of interaction can be referred to as the processing result, and the output of the first model for the remaining rounds of interaction can be referred to as the intermediate interaction results). Taking the aforementioned target user input as "statistical XX data" and including three rounds of interaction as an example, the application management platform 110 may only provide the user with the statistical results of the XX data (i.e., the output of the first model corresponding to the third round of interaction), without providing the user with the intermediate interaction results corresponding to the first and second rounds of interaction.

[0066] In some embodiments, the application management platform 110 can directly determine the processing result output by the first model as the processing result to be presented to the target user. In some embodiments, the application management platform 110 can also process the processing result output by the first model (e.g., add text, formatting, etc.) and determine the processed result as the processing result to be presented to the target user.

[0067] In summary, according to the embodiments of this disclosure, the correlation between multiple candidate interaction records and the current instruction during the interaction process can be determined, and the target interaction record can be determined from the multiple candidate interaction records based on the correlation. The current input for the first model can be determined based on the target interaction record and the current instruction. Thus, context information is filtered based on correlation during the interaction. This allows for dynamic control of the amount of data during the interaction process, avoiding situations where the amount of data is too large. Furthermore, since correlation is considered in the context filtering, this helps to improve the stability of information processing while ensuring the accuracy and efficiency of information processing.

[0068] Embodiments of this disclosure also provide corresponding apparatus for implementing the methods or processes described above. Figure 4 shows a schematic structural block diagram of an apparatus 400 for information processing according to some embodiments of this disclosure. The apparatus 400 may be implemented in or included in an application management platform 110, for example. The various modules / components in the apparatus 400 may be implemented by hardware, software, firmware, or any combination thereof.

[0069] As shown in the figure, the device 400 includes an instruction generation module 410, configured to generate a current instruction for the first model during an interaction between a target user input and the first model. The current instruction indicates the task to be processed by the first model. The device 400 also includes a correlation determination module 420, configured to determine the correlation between multiple candidate interaction records and the current instruction. The multiple candidate interaction records include at least the target user input, one or more historical outputs of the first model during the interaction, or one or more historical instructions for the first model during the interaction. The device 400 also includes a record removal module 430, configured to remove at least one candidate interaction record from the multiple candidate interaction records based on their correlation, to obtain at least one target interaction record. The device 400 also includes an input generation module 440, configured to generate a current input for the first model based on the current instruction and at least one target interaction record, to process the target user input.

[0070] In some embodiments, the correlation determination module 420 is further configured to: determine the correlation between a given interaction record and the current instruction in at least one of the following dimensions for a given interaction record among a plurality of candidate interaction records: time dimension, interaction source dimension, or content dimension; and determine the correlation between the given interaction record and the current instruction based on the correlation in at least one dimension.

[0071] In some embodiments, the first candidate interaction record is generated earlier than the second candidate interaction record, and the correlation between the first candidate interaction record and the second candidate interaction record is lower in the time dimension.

[0072] In some embodiments, the relevance of the target user input to the interaction source dimension is greater than the relevance of one or more historical outputs to the interaction source dimension, and

[0073] One or more historical outputs are more relevant to the interaction source than one or more historical commands are.

[0074] In some embodiments, the relevance of a given interaction record in the content dimension is determined based on the degree of matching between the topic of the given interaction record and the topic of the current instruction.

[0075] In some embodiments, at least one dimension includes multiple dimensions, and the correlation determination module 420 is further configured to: calculate the correlation between a given interaction record and the current instruction based on the corresponding weights of the multiple dimensions and the corresponding correlation of the given interaction record under the multiple dimensions.

[0076] In some embodiments, the apparatus 400 further includes: a topic determination module configured to use a second model to determine a topic input by a target user and a corresponding topic input by one or more historical users prior to the target user input; a matching degree determination module configured to determine a matching degree between the corresponding topic input by one or more historical users and the topic input by the target user; and a record determination module configured to determine historical user inputs with a matching degree greater than a threshold matching degree as candidate interaction records among a plurality of candidate interaction records.

[0077] In some embodiments, the apparatus 400 further includes a quantity determination module configured to determine whether a plurality of candidate interaction records will cause the amount of data in the model input to exceed a threshold data amount, and wherein the operation of removing at least one candidate interaction record from the plurality of candidate interaction records is performed in response to determining that the plurality of candidate interaction records will cause the amount of data in the model input to exceed the threshold data amount, and that the current amount of data in the input is below the threshold data amount.

[0078] In some embodiments, the target user input indicates an operation associated with the data table.

[0079] The units and / or modules included in device 400 can be implemented in various ways, including software, hardware, firmware, or any combination thereof. In some embodiments, one or more units and / or modules can be implemented using software and / or firmware, such as machine-executable instructions stored on a storage medium. In addition to or as an alternative to machine-executable instructions, some or all of the units and / or modules in device 400 can be implemented at least partially by one or more hardware logic components. By way of example and not limitation, exemplary types of hardware logic components that can 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 so on.

[0080] Figure 5 shows a block diagram of an electronic device 500 in which one or more embodiments of the present disclosure may be implemented. It should be understood that the electronic device 500 shown in Figure 5 is merely exemplary and should not constitute any limitation on the functionality and scope of the embodiments described herein. The electronic device 500 shown in Figure 5 may include or be implemented as the application management platform 110 of Figure 1, or the device 400 of Figure 4.

[0081] As shown in Figure 5, the electronic device 500 is in the form of a general-purpose electronic device. Components of the electronic device 500 may include, but are not limited to, one or more processors or processing units 510, memory 520, storage devices 530, one or more communication units 540, one or more input devices 550, and one or more output devices 560. The processing unit 510 may be a physical or virtual processor and is capable of performing various processes according to programs stored in the memory 520. In a multiprocessor system, multiple processing units execute computer-executable instructions in parallel to improve the parallel processing capability of the electronic device 500.

[0082] Electronic device 500 typically includes multiple computer storage media. Such media can be any accessible media that is accessible to electronic device 500, including but not limited to volatile and non-volatile media, removable and non-removable media. Memory 520 can be volatile memory (e.g., registers, cache, random access memory (RAM)), non-volatile memory (e.g., read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. Storage device 530 can be removable or non-removable media and can include machine-readable media, such as flash drives, disks, or any other media that can be used to store information and / or data and can be accessed within electronic device 500.

[0083] Electronic device 500 may further include additional removable / non-removable, volatile / non-volatile storage media. Although not shown in FIG. 5, disk drives for reading from or writing to removable, non-volatile disks (e.g., "floppy disks") and optical disk drives for reading from or writing to removable, non-volatile optical disks may be provided. In these cases, each drive may be connected to a bus (not shown) via one or more data media interfaces. Memory 520 may include computer program product 525 having one or more program modules configured to perform various methods or actions of various embodiments of the present disclosure.

[0084] Communication unit 540 enables communication with other electronic devices via a communication medium. Additionally, the functionality of components of electronic device 500 can be implemented using a single computing cluster or multiple computing machines capable of communicating via communication connections. Therefore, electronic device 500 can operate in a networked environment using logical connections to one or more other servers, network personal computers (PCs), or another network node.

[0085] Input device 550 can be one or more input devices, such as a mouse, keyboard, trackball, etc. Output device 560 can be one or more output devices, such as a monitor, speaker, printer, etc. Electronic device 500 can also communicate with one or more external devices (not shown) via communication unit 540 as needed. These external devices include storage devices, display devices, etc., and can communicate with one or more devices that enable user interaction with electronic device 500, or with any device that enables electronic device 500 to communicate with one or more other electronic devices (e.g., network card, modem, etc.). Such communication can be performed via input / output (I / O) interface (not shown).

[0086] According to an exemplary implementation of this disclosure, a computer-readable storage medium is provided that stores computer-executable instructions thereon, wherein the computer-executable instructions are executed by a processor to implement the methods described above. According to an exemplary implementation of this disclosure, a computer program product is also provided, which is tangibly stored on a non-transitory computer-readable medium and includes computer-executable instructions, which are executed by a processor to implement the methods described above.

[0087] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, devices, and computer program products implemented according to this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0088] These computer-readable program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processing unit of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0089] Computer-readable program instructions can be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions that execute on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0090] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0091] Various implementations of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed implementations. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described implementations. The terminology used herein is chosen to best explain the principles, practical applications, or improvements to technology in the market, or to enable others skilled in the art to understand the various implementations disclosed herein.

Claims

1. An information processing method, comprising: During the interaction between the target user input and the first model, a current instruction is generated for the first model, which indicates the task to be processed by the first model; Determine the correlation degree between multiple candidate interaction records and the current instruction. The multiple candidate interaction records include at least the target user input, one or more historical outputs of the first model during the interaction process, or one or more historical instructions for the first model during the interaction process. Based on the corresponding relevance of the plurality of candidate interaction records, at least one candidate interaction record is removed from the plurality of candidate interaction records to obtain at least one target interaction record; as well as Based on the current instruction and the at least one target interaction record, a current input for the first model is generated to process the target user input.

2. The method according to claim 1, wherein determining the correlation degree between the plurality of candidate interaction records and the current instruction includes: For a given interaction record among the multiple candidate interaction records, Determine the correlation between the given interaction record and the current instruction in at least one of the following dimensions: Time dimension, interaction source dimension, or content dimension; as well as Based on the correlation degree under the at least one dimension, the correlation degree between the given interaction record and the current instruction is determined.

3. The method according to claim 2, wherein the generation time of the first candidate interaction record is earlier than the generation time of the second candidate interaction record, and the correlation degree of the first candidate interaction record in the time dimension is lower than the correlation degree of the second candidate interaction record in the time dimension.

4. The method according to claim 2, wherein the relevance of the target user input under the interaction source dimension is greater than the relevance of the one or more historical outputs under the interaction source, and The correlation between the one or more historical outputs and the interaction source is greater than the correlation between the one or more historical instructions and the interaction source.

5. The method according to claim 2, wherein the relevance of the given interaction record in the content dimension is determined based on the matching degree between the topic of the given interaction record and the topic of the current instruction.

6. The method of claim 2, wherein the at least one dimension comprises multiple dimensions, and determining the correlation between the given interaction record and the current instruction comprises: Based on the corresponding weights of the multiple dimensions and the corresponding correlation of the given interaction record under the multiple dimensions, the correlation between the given interaction record and the current instruction is calculated.

7. The method according to claim 1, further comprising: Using the second model, determine the topic input by the target user and the corresponding topics input by one or more historical users prior to the target user's input; Determine the degree of matching between the topics entered by the one or more historical users and the topics entered by the target user; as well as The historical user inputs with a matching degree greater than the threshold matching degree are determined as candidate interaction records among the multiple candidate interaction records.

8. The method according to claim 1, further comprising: Determine whether the plurality of candidate interaction records will cause the amount of data input to the model to exceed a threshold amount, and The operation of removing at least one candidate interaction record from the plurality of candidate interaction records is performed in response to determining that the plurality of candidate interaction records will cause the amount of data input to the model to exceed the threshold data amount, and that the amount of data input at present is lower than the threshold data amount.

9. The method of claim 1, wherein the target user input indicates an operation associated with the data table.

10. An apparatus for information processing, comprising: The instruction generation module is configured to generate a current instruction for the first model during the interaction between the target user input and the first model, the current instruction indicating the task to be processed by the first model; The correlation determination module is configured to determine the correlation between multiple candidate interaction records and the current instruction. The multiple candidate interaction records include at least the target user input, one or more historical outputs of the first model during the interaction process, or one or more historical instructions for the first model during the interaction process. The record removal module is configured to remove at least one candidate interaction record from the plurality of candidate interaction records based on the corresponding relevance of the plurality of candidate interaction records, so as to obtain at least one target interaction record; as well as An input generation module is configured to generate current input for the first model based on the current instruction and the at least one target interaction record, in order to process the target user input.

11. An electronic device, comprising: At least one processing unit; as well as At least one memory, coupled to the at least one processing unit and storing instructions for execution by the at least one processing unit, which, when executed by the at least one processing unit, cause the electronic device to perform the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, the computer program being executable by a processor to implement the method according to any one of claims 1 to 9.

13. A computer program product comprising a computer program, wherein the computer program, when executed by a processor, implements the method according to any one of claims 1 to 9.

Images