Dialogue recommendation method and device, electronic equipment and storage medium

By splitting the policy selection and policy execution modules in the dialogue recommendation system, and combining intention recognition and sentiment analysis, the problem of low recommendation accuracy in the existing technology is solved, and more efficient personalized recommendation is achieved.

CN119088925APending Publication Date: 2024-12-06PICC INFORMATION TECH CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411183394.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-26
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing dialogue recommendation system uses fixed rules to recommend, resulting in low recommendation accuracy and failing to effectively meet users' personalized needs.

Method used

Through the splitting of the policy selection module and the policy execution module, personalized recommendation strategies are generated and executed, and combined with the intention identification module and sentiment analysis algorithm, the accuracy of recommendation information is optimized.

Benefits of technology

It improves the relevance and accuracy of recommended information, can better meet users' personalized needs and improve user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119088925A_ABST
    Figure CN119088925A_ABST
Patent Text Reader

Abstract

The invention provides a dialogue recommendation method and device, electronic equipment and a storage medium. The method comprises the following steps: receiving a recommendation request, wherein the recommendation request comprises initial input information; according to the recommendation request, a target recommendation strategy and a target recommendation type corresponding to the initial input information are determined through a strategy selection module, and the strategy selection module comprises a plurality of sub-modules corresponding to a plurality of recommendation types; and according to the target recommendation type, executing the target recommendation strategy through a strategy execution module to obtain target recommendation information. According to the scheme, the recommendation strategy is generated through the strategy selection module, the recommendation strategy is executed through the strategy execution module, the strategy selection module and the strategy execution module can focus on respective functions, function confusion is avoided, and therefore recommendation accuracy is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a conversation recommendation method, device, electronic device and storage medium. Background Art

[0002] The conversational recommendation system is an application that combines conversation management and personalized recommendation technology. The conversational recommendation system can understand the user's needs by interacting with the user, determine customized content based on the user's needs, and recommend the customized content to the user.

[0003] In the related art, the dialogue recommendation system includes a general decision module, which can analyze user needs and determine corresponding recommendation information.

[0004] However, general decision modules make recommendations based on fixed rules, which results in low recommendation accuracy. Summary of the invention

[0005] The present application provides a conversation recommendation method, device, electronic device and storage medium to improve the accuracy of recommendations.

[0006] In a first aspect, the present application provides a dialogue recommendation method, comprising: receiving a recommendation request, the recommendation request comprising initial input information; determining, based on the recommendation request, a target recommendation strategy and a target recommendation type corresponding to the initial input information through a strategy selection module, the strategy selection module comprising multiple sub-modules corresponding to multiple recommendation types; based on the target recommendation type, executing the target recommendation strategy through a strategy execution module to obtain target recommendation information.

[0007] In one possible implementation, a target recommendation strategy and a target recommendation type corresponding to the initial input information are determined through a strategy selection module, including: determining the target information type of the initial input information, the target information type including text or voice; if the target information type is text, determining the initial input information as intermediate input information; if the target information type is voice, processing the initial information through voice recognition technology to obtain intermediate input information; determining the target recommendation strategy and the target recommendation type corresponding to the intermediate input information through a strategy selection module.

[0008] In one possible implementation, the strategy selection module includes an intent recognition module; determining the target recommendation strategy and target recommendation type corresponding to the intermediate input information through the strategy selection module, including: performing intent recognition processing on the intermediate input information through the intent recognition module to obtain the target recommendation type; determining the target sub-module corresponding to the target recommendation type; processing the intermediate input information through the target sub-module to obtain the target recommendation strategy.

[0009] In a possible implementation, the target recommendation strategy is executed according to the target recommendation type by a strategy execution module to obtain target recommendation information, including: determining a recommendation model, wherein the recommendation model is obtained by training a training sample; and inputting the target recommendation type and the target recommendation strategy into the recommendation model to obtain the target recommendation information.

[0010] In a possible implementation, the method further includes: determining a recommendation window, the recommendation window being used to store the initial input information and the target recommendation information, the recommendation window being established based on the recommendation request; determining a target number of target recommendation information in the recommendation window; and closing the recommendation window if a stop recommendation indication is received and / or the target number is greater than or equal to a quantity threshold.

[0011] In a possible implementation, the method further includes: receiving multimodal feedback information for the target recommendation information, the multimodal feedback information including multiple information types; analyzing and processing the multimodal feedback information through a sentiment analysis algorithm model to obtain target emotion information; and adjusting the output information of the strategy selection module and the strategy execution module respectively through the target emotion information.

[0012] In a possible implementation, the recommendation request also includes an output information type, and the output information type includes text and / or voice; the method also includes: if the output information type is text, converting the target recommendation information into text content; if the output information type is voice, converting the target recommendation information into voice content; if the output information type is text and voice, converting the target recommendation information into text content and language content respectively.

[0013] In a second aspect, the present application provides a dialogue recommendation device, comprising: a receiving module for receiving a recommendation request, the recommendation request comprising initial input information; a determination module for determining, based on the recommendation request, a target recommendation strategy and a target recommendation type corresponding to the initial input information through a strategy selection module, the strategy selection module comprising multiple sub-modules corresponding to multiple recommendation types; an execution module for executing the target recommendation strategy through a strategy execution module according to the target recommendation type to obtain target recommendation information.

[0014] In a possible implementation, the determination module is specifically used to determine the target information type of the initial input information, and the target information type includes text or voice; the determination module is also specifically used to determine the initial input information as intermediate input information if the target information type is text; the determination module is also specifically used to process the initial information through voice recognition technology to obtain intermediate input information if the target information type is voice; the determination module is also specifically used to determine the target recommendation strategy and target recommendation type corresponding to the intermediate input information through a strategy selection module.

[0015] In one possible implementation, the strategy selection module includes an intention recognition module; the determination module is specifically used to perform intent recognition processing on the intermediate input information through the intention recognition module to obtain the target recommendation type; the determination module is specifically used to determine the target sub-module corresponding to the target recommendation type; the determination module is specifically used to process the intermediate input information through the target sub-module to obtain the target recommendation strategy.

[0016] In a possible implementation, the device also includes: a query module, used to determine a recommendation model, wherein the recommendation model is obtained by performing model training on training samples; the query module is also used to input the target recommendation type and the target recommendation strategy into the recommendation model to obtain the target recommendation information.

[0017] In a possible implementation, the device also includes: a window module, used to determine a recommendation window, the recommendation window is used to store the initial input information and the target recommendation information, the recommendation window is established according to the recommendation request; the window module is also used to determine the target number of target recommendation information in the recommendation window; the window module is also used to close the recommendation window if a stop recommendation indication is received and / or the target number is greater than or equal to a quantity threshold.

[0018] In a possible implementation, the device also includes: an adjustment module, used to receive multimodal feedback information for the target recommendation information, the multimodal feedback information including multiple information types; the adjustment module is also used to analyze and process the multimodal feedback information through a sentiment analysis algorithm model to obtain target emotional information; the adjustment module is also used to adjust the output information of the strategy selection module and the strategy execution module respectively according to the target emotional information.

[0019] In a possible implementation, the recommendation request also includes an output information type, and the output information type includes text and / or voice; the device also includes: an output module, which is used to convert the target recommendation information into text content if the output information type is text; the output module is also used to convert the target recommendation information into voice content if the output information type is voice; the output module is also used to convert the target recommendation information into text content and language content respectively if the output information type is text and voice.

[0020] In a third aspect, the present application provides an electronic device comprising: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement any method described in the first aspect.

[0021] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are executed by a processor as described in any one of the methods in the first aspect.

[0022] In a fifth aspect, the present application provides a computer program product, comprising a computer program, wherein the computer program is executed by a processor as in any one of the methods in the first aspect.

[0023] The present application provides a method, device, electronic device and storage medium for dialogue recommendation, the method comprising: receiving a recommendation request, the recommendation request comprising initial input information; according to the recommendation request, determining the target recommendation strategy and target recommendation type corresponding to the initial input information through a strategy selection module, the strategy selection module comprising multiple submodules corresponding to multiple recommendation types; according to the target recommendation type, executing the target recommendation strategy through a strategy execution module to obtain target recommendation information. In the above scheme, the recommendation strategy is generated through the strategy selection module and executed through the strategy execution module. The strategy selection module and the strategy execution module can focus on their respective functions to avoid functional confusion, thereby improving the accuracy of recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0025] Figure 1 A schematic diagram of an application scenario of a conversation recommendation method provided in an embodiment of the present application;

[0026] Figure 2 A flowchart of a conversation recommendation method provided in an embodiment of the present application;

[0027] Figure 3 A flowchart of a conversation recommendation method provided in an embodiment of the present application;

[0028] Figure 4 A schematic diagram of a target determination submodule provided in an embodiment of the present application;

[0029] Figure 5 A schematic diagram of a target recommendation strategy for generating a target provided in an embodiment of the present application;

[0030] Figure 6 A schematic diagram of optimization based on feedback information provided in an embodiment of the present application;

[0031] Figure 7 A schematic diagram of the structure of a conversation recommendation device provided in an embodiment of the present application;

[0032] Figure 8 A schematic diagram of the structure of a conversation recommendation device provided in an embodiment of the present application;

[0033] Fig. 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0034] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0035] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0036] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, processing, transmission, provision, disclosure and application of the relevant data comply with the relevant laws, regulations and standards of the relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0037] In addition, this application involves big data analysis of user information (including but not limited to personal biometrics, identity data, consumption data, asset data, electronic terminal operation data, etc.), and the use of artificial intelligence technology for automated decision-making, and a technical solution for making decisions that have a significant impact on personal rights and interests based on the results of automated decision-making. The application provides users with corresponding operation entrances for them to choose to agree or reject the results of automated decision-making; if the user chooses to reject, the expert decision-making process will be entered.

[0038] It should be noted that the conversation recommendation method, device, electronic device and storage medium of the present application can be used in the field of artificial intelligence technology, and can also be used in any field other than artificial intelligence. The application field of the conversation recommendation method, device, electronic device and storage medium of the present application is not limited.

[0039] Figure 1 A schematic diagram of an application scenario of a conversation recommendation method provided in an embodiment of the present application is used as an example in combination with the scenario shown in the figure: a user operates a terminal device, and the terminal device sends the user input content as input information to a conversation recommendation system. The conversation recommendation system sends recommendation information corresponding to the input information to the terminal device.

[0040] Combined with the scenario example, the conversational recommendation focuses on analyzing the user's needs through interaction with the user and determining the recommended information corresponding to the needs. For example, the recommended information includes but is not limited to web pages, movies, or music.

[0041] In the related art, the recommended information is determined by a general decision module. However, in actual applications, the input information usually includes the personalized needs of users, and the rules in the general decision module cannot fully consider the personalized needs, resulting in the lack of personalization of the recommended information and failure to accurately meet the needs of users.

[0042] The method for dialogue recommendation provided in this application is intended to solve the above technical problems in the prior art.

[0043] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0044] Figure 2 A flowchart of a conversation recommendation method provided in an embodiment of the present application, the method comprising the following steps:

[0045] S201: Receive a recommendation request, where the recommendation request includes initial input information.

[0046] As an example, the execution subject of this embodiment may be a conversation recommendation device, and there are many ways to implement the conversation recommendation device. For example, it may be program software, or a medium storing relevant computer programs, such as a USB flash drive, etc.; or, the device may also be a physical device integrated or installed with relevant computer programs, such as a chip, a smart terminal, a computer, a server, etc.

[0047] The initial input information is original information input by the user, and the initial input information describes the user's needs.

[0048] In conjunction with the scenario example, the recommendation request is used to instruct the dialogue recommendation system to generate recommendation information corresponding to the initial input information.

[0049] S202. According to the recommendation request, a target recommendation strategy and a target recommendation type corresponding to the initial input information are determined by a strategy selection module. The strategy selection module includes a plurality of submodules corresponding to a plurality of recommendation types.

[0050] Among them, this application divides the dialogue recommendation system into a strategy selection module and a strategy execution module. The strategy selection module is used to understand the user's needs, intentions, or context information, and decide which type of strategy to use to respond to the recommendation request. The strategy execution module is used to specifically execute the strategy to obtain the target recommendation information.

[0051] In the related art, the dialogue recommendation system includes a general decision-making module, in which the strategy selection and strategy execution functions are highly coupled.

[0052] Exemplarily, the recommendation type includes but is not limited to at least one of the following: recommended content, chat interaction, or preference attribute mining, etc.

[0053] It can be understood that by splitting into a policy selection module and a policy execution module, the decoupling of the policy selection and policy execution functions can be achieved, thereby effectively splitting the policy selection module into multiple sub-modules.

[0054] Furthermore, multiple recommendation types corresponding to multiple sub-modules can capture the user's personalized needs in more detail for multiple dimensions, thereby improving the relevance of recommended information.

[0055] S203. According to the target recommendation type, the target recommendation strategy is executed by the strategy execution module to obtain target recommendation information.

[0056] Exemplarily, the strategy execution module executes the target recommendation strategy in a targeted manner according to the target recommendation type.

[0057] Combined with the scenario example, if the target recommendation type is recommended content, the content is filtered according to the user's historical preferences and the current conversation content. If the target recommendation type is chat interaction, a non-target-oriented communication with the user is generated. If the target recommendation type is preference attribute mining, questions are generated to understand user needs in more detail.

[0058] The dialogue recommendation method provided by the embodiment of the present application includes: receiving a recommendation request, the recommendation request includes initial input information; according to the recommendation request, determining the target recommendation strategy and target recommendation type corresponding to the initial input information through a strategy selection module, the strategy selection module includes multiple submodules corresponding to multiple recommendation types; according to the target recommendation type, executing the target recommendation strategy through a strategy execution module to obtain target recommendation information. In the above scheme, the recommendation strategy is generated through the strategy selection module and executed through the strategy execution module. The strategy selection module and the strategy execution module can focus on their respective functions to avoid functional confusion, thereby improving the accuracy of recommendations.

[0059] Based on any of the above embodiments, Figure 3 , the detailed process of dialogue recommendation is explained.

[0060] Figure 3 A flowchart of a conversation recommendation method provided in an embodiment of the present application is shown below. Figure 3 As shown, the method includes:

[0061] S301: Receive a recommendation request, where the recommendation request includes initial input information.

[0062] It should be noted that the execution process of S301 refers to S201 and will not be repeated here.

[0063] S302: Determine the target information type of the initial input information.

[0064] The target information type includes text or voice.

[0065] Optionally, the user interface includes controls for text and language, and the controls are used to determine the target information type.

[0066] In conjunction with the scenario example, the target information type is in the form of initial input information, and the user can choose to input it through text or language.

[0067] S303: If the target information type is text, the initial input information is determined as the intermediate input information.

[0068] Combined with the scenario example, the intent recognition module in this application is built based on the language model. The intent recognition module has a text recognition function, and the intermediate input information is used to input the intent recognition module. Therefore, the initial input information of the text type does not need further processing.

[0069] S304: If the target information type is speech, the initial information is processed by speech recognition technology to obtain intermediate input information.

[0070] Among them, speech recognition technology (Automatic Speech Recognition, referred to as ASR) is used to convert speech into text, including but not limited to the following steps: language signal processing, feature extraction, acoustic modeling, language modeling, decoding, or post-processing.

[0071] Combined with the scenario example, the intent recognition module cannot directly recognize speech. Converting speech into text through speech recognition technology can enable the intent module to recognize accurately.

[0072] S305: Perform intent recognition processing on the intermediate input information through the intent recognition module to obtain the target recommendation type.

[0073] Among them, the strategy selection module includes an intention recognition module.

[0074] Exemplarily, the intent recognition module is used to determine a recommended type of input information.

[0075] Optionally, the intent recognition module is built based on a Bert model, which is obtained through unsupervised training of a large amount of unlabeled text. The Bert model can process text from front to back and from back to front at the same time, so as to better understand the relationship between words in the text.

[0076] Combined with the scenario example, by determining the target recommendation type, targeted recommendation processing can be carried out to improve the accuracy of recommendations.

[0077] S306: Determine the target submodule corresponding to the target recommendation type.

[0078] Optionally, a correspondence between the recommendation type and the submodule is determined, and a target submodule is determined according to the target recommendation type and the correspondence.

[0079] Exemplarily, the strategy selection module is split into multiple sub-modules, each sub-module is used to process a corresponding recommendation type.

[0080] Next, combine Figure 4 The target determination submodule is described.

[0081] Figure 4 Schematic diagram of the target determination submodule provided in the embodiment of the present application. Figure 4 As shown, the intermediate input information is processed for intent recognition to obtain the target recommendation type corresponding to the intermediate input information. The corresponding relationship between multiple recommendation types and multiple sub-modules is determined. The target sub-module is determined according to the target recommendation type and the corresponding relationship.

[0082] S307: Process the intermediate input information through the target submodule to obtain a target recommendation strategy.

[0083] Exemplarily, the target submodule determines information such as the user's behavior, preference, or context based on the intermediate input information, and generates a target recommendation strategy based on the information.

[0084] Optionally, before generating the target recommendation strategy, the target submodule performs data preprocessing on the intermediate input information, and the data preprocessing includes but is not limited to at least one of the following: data cleaning, data standardization, or missing value processing, etc. It can be understood that the content of the intermediate input information can be accurately understood through data preprocessing, thereby improving the accuracy of the target recommendation information.

[0085] Next, combine Figure 5 The strategy for generating target recommendations is described.

[0086] Figure 5 Schematic diagram of a target recommendation strategy for generating a target recommendation strategy provided in an embodiment of the present application. Figure 5 As shown, during the execution of the strategy selection module, the intent recognition module is used to recognize the intent of the initial input information, and a target submodule is determined from multiple submodules based on the intent recognition result. A target recommendation strategy is generated through the target submodule, and the target recommendation strategy is output.

[0087] S308: Determine a recommendation model, where the recommendation model is obtained by performing model training on training samples.

[0088] Exemplarily, historical input information is obtained, the historical input information is used as a training sample, a machine learning algorithm is determined, and a model training is performed on the historical input information through the machine learning algorithm to obtain a recommendation model.

[0089] Optionally, the machine learning algorithm may be clustering, logistic regression, support vector machine, random forest, or neural network.

[0090] Optionally, the historical input information covers multiple recommendation types.

[0091] Combined with the scenario examples, the recommendation model obtained by training the model through multiple types of historical input information can learn the business scenarios corresponding to each recommendation type, thereby generating accurate recommendation information for each recommendation type.

[0092] S309: Input the target recommendation type and target recommendation strategy into the recommendation model to obtain target recommendation information.

[0093] Exemplarily, the recommendation model first determines the scope of the recommendation information according to the target recommendation type, and then executes the target recommendation strategy according to the scope to obtain the target recommendation information.

[0094] Combined with the scenario example, when the target recommendation type is recommended content, the recommended items are selected based on the user's historical preferences and the current conversation content. In this process, the recommendation model can be used to enhance the recommendation effect. The recommendation model extracts key features from the user's historical behavior, such as past browsing history, purchase history, search keywords, etc. The recommendation model understands the content of the input information to capture the user's immediate interest. On this basis, it predicts the content that the user may be interested in. When the target recommendation type is chat interaction, the conversational recommendation system needs to have non-goal-oriented communication with the user, which is usually to build trust or keep the user active. On this basis, the recommendation model generates natural and smooth conversation responses. The recommendation model is also used to monitor the emotional tendency of the conversation to ensure that the communication is positive. The recommendation model is also used to maintain the coherence and fun of the conversation and encourage users to continue to participate. When the target recommendation type is preference attribute mining, the conversational recommendation system needs to generate questions to understand the user's needs and preferences in more detail in order to better provide services to the user. According to the user's needs, the recommendation model generates relevant questions, sorts the questions according to importance and relevance, and ensures that the most critical questions are asked first. The recommendation model is also used to adjust subsequent questions or recommendation strategies based on the user's answers.

[0095] In a feasible implementation method, the dialogue recommendation method also includes: determining a recommendation window, the recommendation window is used to store initial input information and target recommendation information, and the recommendation window is established according to the recommendation request; determining the target number of target recommendation information in the recommendation window; if a stop recommendation instruction is received and / or the target number is greater than or equal to a quantity threshold, closing the recommendation window.

[0096] Among them, the user's input information and the recommended information of the dialogue recommendation system are all displayed in the recommendation window. The user can trace the historical dialogue records through the window to improve the efficiency of dialogue recommendations.

[0097] Exemplarily, after the current recommendation window is closed, if a recommendation request is received again, a new recommendation window is created, and the recommendation information is displayed in the new recommendation window.

[0098] Combined with the scenario example, if the number of conversation turns is too long, it will increase the load of the conversation recommendation system, reduce the operating efficiency of the conversation recommendation system, and affect the user experience.

[0099] In this feasible implementation, closing the recommendation window when the rule is met can avoid excessive load on the dialogue recommendation system, thereby improving the execution efficiency of the dialogue recommendation system.

[0100] In a feasible implementation method, the dialogue recommendation method also includes: receiving multimodal feedback information for target recommendation information, the multimodal feedback information includes multiple information types; analyzing and processing the multimodal feedback information through a sentiment analysis algorithm model to obtain target emotional information; and adjusting the output information of the strategy selection module and the strategy execution module respectively through the target emotional information.

[0101] Among them, the feedback information is used to optimize the recommendation information.

[0102] Optionally, the feedback information may be in the form of text, voice, or video.

[0103] Combined with the scenario example, the feedback information includes the user's emotional state or emotional fluctuations. For example, the user's tone, wording or expression in the feedback information can reflect the user's satisfaction with the recommended information. The recommended information is optimized according to the satisfaction level until the recommended information that satisfies the user is obtained. It can be understood that multimodal sentiment analysis can effectively improve the analysis accuracy compared to single-modal sentiment analysis.

[0104] Exemplarily, the model is iteratively optimized through feedback information to improve the output accuracy of the iteratively optimized model.

[0105] Next, combine Figure 6 Optimization based on feedback information is described.

[0106] Figure 6 Schematic diagram of optimization based on feedback information provided in the embodiment of the present application. Figure 6 As shown, the intermediate input information is determined according to the information type of the initial input information, and the intermediate input information is processed by the strategy selection module and the strategy execution module to obtain the target recommendation information. Feedback information for the target recommendation information is received. If the feedback information meets the termination condition, the recommendation is terminated. If the termination condition is not met, the target recommendation information is regenerated through the multimodal sentiment analysis results until the termination condition is met.

[0107] In this feasible implementation, through multimodal sentiment analysis, the user's emotional expression can be fully captured, the error of sentiment analysis can be reduced, and the accuracy of target recommendation information can be improved.

[0108] In a feasible implementation method, the recommendation request also includes an output information type, and the output information type includes text and / or voice; the dialogue recommendation method also includes: if the output information type is text, converting the target recommendation information into text content; if the output information type is voice, converting the target recommendation information into voice content; if the output information type is text and voice, converting the target recommendation information into text content and language content respectively.

[0109] Exemplarily, in the user interface, the user can select the output information type, and the dialogue recommendation system generates corresponding content based on the user's selection.

[0110] In combination with the scenario example, the user determines the output information type according to the current scenario. For example, if the user is currently in a place where quietness is required, such as a library, the output information type is selected as text.

[0111] In this feasible implementation, corresponding recommendation information is generated according to the output information type selected by the user, which is in line with the user's current application scenario, thereby prompting the user's usage experience.

[0112] Figure 7 This is a schematic diagram of the structure of a conversation recommendation device provided in an embodiment of the present application. Figure 7 As shown, the dialogue recommendation device 70 may include: a receiving module 71, a determining module 72 and an executing module 73, wherein:

[0113] The receiving module 71 is used to receive a recommendation request, where the recommendation request includes initial input information.

[0114] The determination module 72 is used to determine the target recommendation strategy and target recommendation type corresponding to the initial input information through the strategy selection module according to the recommendation request. The strategy selection module includes multiple sub-modules corresponding to multiple recommendation types.

[0115] The execution module 73 is used to execute the target recommendation strategy through the strategy execution module according to the target recommendation type to obtain the target recommendation information.

[0116] Optionally, the receiving module 71 may execute Figure 2 S201 in the embodiment.

[0117] Optionally, the determination module 72 may execute Figure 2 S202 in the embodiment.

[0118] Optionally, the execution module 73 may execute Figure 2S203 in the embodiment.

[0119] It should be noted that the dialogue recommendation device shown in the embodiment of the present application can execute the technical solution shown in the above method embodiment, and its implementation principle and beneficial effects are similar, which will not be repeated here.

[0120] In a possible implementation, the determination module 72 is specifically configured to:

[0121] Determine the target information type of the initial input information, which includes text or voice; if the target information type is text, determine the initial input information as intermediate input information; if the target information type is voice, process the initial information through voice recognition technology to obtain intermediate input information; determine the target recommendation strategy and target recommendation type corresponding to the intermediate input information through the strategy selection module.

[0122] In a possible implementation, the strategy selection module includes an intention recognition module; a determination module 72, specifically configured to:

[0123] The intermediate input information is processed for intent recognition through the intent recognition module to obtain the target recommendation type; the target sub-module corresponding to the target recommendation type is determined; the intermediate input information is processed through the target sub-module to obtain the target recommendation strategy.

[0124] Figure 8 A schematic diagram of the structure of a conversation recommendation device provided in an embodiment of the present application. Figure 7 Based on the embodiment shown, Figure 8 As shown, the dialogue recommendation device 80 further includes: a query module 74, a window module 75, an adjustment module 76 and an output module 77, wherein:

[0125] The query module 74 is used to:

[0126] Determine the recommendation model, which is obtained by training the training samples; input the target recommendation type and target recommendation strategy into the recommendation model to obtain the target recommendation information.

[0127] The window module 75 is used for:

[0128] Determine a recommendation window, where the recommendation window is used to store initial input information and target recommendation information, and the recommendation window is established based on a recommendation request; determine a target number of target recommendation information in the recommendation window; and close the recommendation window if a stop recommendation instruction is received and / or the target number is greater than or equal to a quantity threshold.

[0129] The adjustment module 76 is used to:

[0130] Receive multimodal feedback information for target recommendation information, the multimodal feedback information including multiple information types; analyze and process the multimodal feedback information through a sentiment analysis algorithm model to obtain target sentiment information; and adjust the output information of a strategy selection module and a strategy execution module respectively through the target sentiment information.

[0131] The recommendation request also includes an output information type, which includes text and / or voice; an output module 77, which is used to:

[0132] If the output information type is text, the target recommended information is converted into text content; if the output information type is voice, the target recommended information is converted into voice content; if the output information type is text and voice, the target recommended information is converted into text content and language content respectively.

[0133] Fig. 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown in FIG. Fig. 9 As shown, the electronic device includes:

[0134] The electronic device includes a processor 291 and a memory 292; it may also include a communication interface 293 and a bus 294. The processor 291, the memory 292, and the communication interface 293 may communicate with each other through the bus 294. The communication interface 293 may be used for information transmission. The processor 291 may call the logic instructions in the memory 292 to execute the method of the above embodiment.

[0135] In addition, the logic instructions in the above-mentioned memory 292 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.

[0136] The memory 292 is a computer-readable storage medium that can be used to store software programs and computer executable programs, such as program instructions / modules corresponding to the methods in the embodiments of the present application. The processor 291 executes functional applications and data processing by running the software programs, instructions, and modules stored in the memory 292, that is, implementing the methods in the above method embodiments.

[0137] The memory 292 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function; the data storage area may store data created according to the use of the terminal device, etc. In addition, the memory 292 may include a high-speed random access memory and may also include a non-volatile memory.

[0138] An embodiment of the present application provides a non-temporary computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the method as described in the above embodiment.

[0139] An embodiment of the present application provides a computer program product, including a computer program, which implements the method of the above embodiment when the computer program is executed by a processor.

[0140] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0141] It should be further noted that, although the various steps in the flowchart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowchart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0142] It should be understood that the above-mentioned device embodiments are only illustrative, and the device of the present application can also be implemented in other ways. For example, the division of units / modules in the above-mentioned embodiments is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units, modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0143] In addition, unless otherwise specified, each functional unit / module in each embodiment of the present application may be integrated into one unit / module, each unit / module may exist physically separately, or two or more units / modules may be integrated together. The above-mentioned integrated unit / module may be implemented in the form of hardware or in the form of a software program module.

[0144] If the integrated unit / module is implemented in the form of hardware, the hardware may be a digital circuit, an analog circuit, etc. The physical implementation of the hardware structure includes but is not limited to transistors, memristors, etc. The processor may be any appropriate hardware processor, such as a CPU, a GPU, an FPGA, a DSP, and an ASIC, etc. The storage unit may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc.

[0145] If the integrated unit / module is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0146] In the above embodiments, the description of each embodiment has its own emphasis. For the part not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0147] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0148] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A conversation recommendation method, characterized in that: include: receiving a recommendation request, the recommendation request including initial input information; According to the recommendation request, determining a target recommendation strategy and a target recommendation type corresponding to the initial input information through a strategy selection module, wherein the strategy selection module includes a plurality of submodules corresponding to a plurality of recommendation types; According to the target recommendation type, the target recommendation strategy is executed by a strategy execution module to obtain target recommendation information.

2. The method according to claim 1, characterized in that Determining the target recommendation strategy and target recommendation type corresponding to the initial input information through the strategy selection module includes: Determining a target information type of the initial input information, the target information type including text or voice; If the target information type is text, determining the initial input information as intermediate input information; If the target information type is speech, the initial information is processed by speech recognition technology to obtain intermediate input information; The target recommendation strategy and target recommendation type corresponding to the intermediate input information are determined through a strategy selection module.

3. The method according to claim 2, characterized in that The strategy selection module includes an intention recognition module; Determining the target recommendation strategy and target recommendation type corresponding to the intermediate input information through the strategy selection module includes: Performing intent recognition processing on the intermediate input information by the intent recognition module to obtain the target recommendation type; Determine a target submodule corresponding to the target recommendation type; The intermediate input information is processed by the target submodule to obtain the target recommendation strategy.

4. The method according to any one of claims 1 to 3, characterized in that Executing the target recommendation strategy according to the target recommendation type through a strategy execution module to obtain target recommendation information, including: Determine a recommendation model, where the recommendation model is obtained by performing model training on training samples; The target recommendation type and the target recommendation strategy are input into the recommendation model to obtain the target recommendation information.

5. The method according to claim 4, characterized in that The method further comprises: determining a recommendation window, wherein the recommendation window is used to store the initial input information and the target recommendation information, and the recommendation window is established according to the recommendation request; Determine a target number of target recommendation information in the recommendation window; If a stop recommendation instruction is received and / or the target quantity is greater than or equal to a quantity threshold, the recommendation window is closed.

6. The method according to any one of claims 1 to 5, characterized in that The method further comprises: receiving multimodal feedback information for the target recommendation information, wherein the multimodal feedback information includes multiple information types; Analyzing and processing the multimodal feedback information through a sentiment analysis algorithm model to obtain target sentiment information; The output information of the strategy selection module and the output information of the strategy execution module are adjusted respectively according to the target emotion information.

7. The method according to any one of claims 1 to 6, characterized in that The recommendation request also includes an output information type, and the output information type includes text and / or voice; the method further includes: If the output information type is text, converting the target recommendation information into text content; If the output information type is voice, converting the target recommendation information into voice content; If the output information type is text and voice, the target recommendation information is converted into text content and language content respectively.

8. A dialogue recommendation device, characterized in that: include: A receiving module, configured to receive a recommendation request, wherein the recommendation request includes initial input information; A determination module, configured to determine, according to the recommendation request, a target recommendation strategy and a target recommendation type corresponding to the initial input information through a strategy selection module, wherein the strategy selection module includes a plurality of submodules corresponding to a plurality of recommendation types; The execution module is used to execute the target recommendation strategy through the strategy execution module according to the target recommendation type to obtain target recommendation information.

9. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

Citation Information

Patent Citations

  • Recommendation system and method, electronic equipment and computer readable storage medium

    CN111966908A

  • Information recommendation method and device, equipment and medium

    CN112182047A

  • Information recommendation method and device, storage medium and electronic equipment

    CN116304007A

  • Scientific and technological information push service system based on big data

    CN118445479A

  • Method, apparatus, device, and storage medium for intention recommendation

    US20210209109A1