Information processing method and device, computer equipment, storage medium and program product

Abstract

The invention relates to artificial intelligence, and provides an information processing method and device, computer equipment, a storage medium and a program product. The method comprises the following steps: generating target prompt information for a session text in a first time period; obtaining a reply text responding to the target prompt information; scoring based on the session text and the reply text to obtain a first score of the first user in the first time period; a reference score is determined, evaluation information of the first user is generated according to the first score and the reference score, and the reference score comprises at least one of a second score of the first user in the second time period and a third score of the second user. The method can improve the accuracy of the evaluation result.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and specifically to an information processing method, apparatus, computer equipment, storage medium and program product. Background Art

[0002] In the related art, after the user completes the communication with the agent or intelligent robot, the artificial intelligence algorithm is used to score the communication content between the user and the agent or intelligent robot. However, in the process of analyzing the communication content, the user's real situation cannot be fully measured, resulting in low accuracy of the evaluation results. Summary of the invention

[0003] The present application provides an information processing method, apparatus, computer equipment, storage medium and program product to solve the technical problem that the overall evaluation of users cannot be comprehensively measured.

[0004] A first aspect of an embodiment of the present application provides an information processing method, the method comprising: generating target prompt information for a conversation text within a first time period; obtaining a reply text in response to the target prompt information; scoring the conversation text and the reply text to obtain a first score of a first user in the first time period; determining a reference score, and generating evaluation information for the first user based on the first score and the reference score, the reference score comprising at least one of a second score of the first user in the second time period and a third score of the second user.

[0005] According to a second aspect of an embodiment of the present application, there is provided an information processing device, the device comprising: a generation unit, configured to generate target prompt information for a conversation text within a first time period; an acquisition unit, configured to acquire a reply text in response to the target prompt information; a scoring unit, configured to score the conversation text and the reply text to obtain a first score of a first user in the first time period; the generation unit, further configured to determine a reference score, and generate evaluation information for the first user based on the first score and the reference score, the reference score comprising at least one of a second score of the first user in the second time period and a third score of the second user.

[0006] A third aspect of an embodiment of the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps in the method provided in the first aspect when executing the computer program.

[0007] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the method provided in the first aspect are implemented.

[0008] A fifth aspect of an embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps in the method provided in the first aspect are implemented.

[0009] In the information processing method of this embodiment, by generating target prompt information from the conversation text, the user's reply to the conversation text can be guided, and the text matching degree between the reply text and the conversation text can be improved, thereby assisting the interaction between the user and the intelligent robot, improving the text matching degree between the reply text and the conversation text, and helping to improve the accuracy of the first score based on the two. Further, by combining the first score and the reference score to generate evaluation information, the reference score may include the second score of the first user in the second time period, and may also include the third score of the second user, or both. It can be seen that the evaluation information integrates multiple scores, so that the generated evaluation information can comprehensively measure the real situation of the user and improve the accuracy of the evaluation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0011] Figure 1 It is a schematic diagram of an application scenario of an information processing method provided in an embodiment of the present application;

[0012] Figure 2 is a flow chart of an information processing method provided by an embodiment of the present application;

[0013] Figure 3A , Figure 3B , Figure 3C It is a schematic diagram of a dialogue interface provided in an embodiment of the present application;

[0014] Figure 4 It is a network structure diagram of an information extraction model provided by an embodiment of the present application;

[0015] Figure 5 This is a flow chart of updating a template library provided by an embodiment of the present application;

[0016] Figure 6is a flow chart of updating the information base provided by an embodiment of the present application;

[0017] Figure 7 is a flowchart of calculating a first score provided by an embodiment of the present application;

[0018] Figure 8 is a flowchart of another information processing method provided by an embodiment of the present application;

[0019] Fig. 9 is a functional module diagram of an information processing device provided in an embodiment of the present application;

[0020] Fig.10 It is a structural diagram of a computer device for implementing an information processing method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field in this application. The terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. It should be understood that, unless otherwise specified in this application, " / " means or. For example, A / B can represent A or B. "And / or" in this application is only a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. "At least one" means one or more. "Multiple" means two or more than two. For example, at least one of a, b or c can represent: a, b, c, a and b, a and c, b and c, a, b and c seven situations.

[0024] Some terminology explanation:

[0025] Natural Language Processing (NLP): NLP is an important direction in the fields of computer science and artificial intelligence. It studies various theories and methods that can achieve effective communication between humans and computers using natural language. At present, NLP theoretical research and model exploration have gradually expanded to a wider range of fields that are closer to human intelligence. The problems involved have also gradually extended from tasks such as word segmentation, semantic classification and matching, and question answering, showing a trend from easy to difficult.

[0026] Prompt Learning: It is a natural language generation technology based on human prompts. It adds additional text to the input segment to better use the knowledge of the pre-trained language model, aiming to improve the generation quality and efficiency of the final generated model. The basic idea is to give a human prompt or a partial generation result, called a prompt token, during the generation process, so that the model can continue to generate text on this basis. Prompt learning can be applied to various natural language generation tasks, such as text summarization, machine translation, dialogue generation, etc. Prompt learning can help the model better understand the context and context of the generation task, so as to generate more accurate and natural text, which can improve the generation quality and efficiency of the generation model, and also reduce the generation space of the generation model, thereby improving the generation efficiency and speed.

[0027] At present, in the field of person-job matching, job seekers are evaluated only based on the communication between them and the interviewer during the interview process. This often fails to obtain comprehensive information about the job seekers, making it impossible to assist job seekers in quickly finding a job that matches them, and is also not conducive to recruitment by employers.

[0028] In training fields such as pre-job training and language training, since the user's learning situation is evaluated only by the user's single test results, it is often impossible to fully measure the user's actual learning situation, resulting in low accuracy of the evaluation results.

[0029] Based on this, the embodiment of the present application provides an information processing method, which generates target prompt information through conversation text, can guide users to reply to conversation text, improve the text matching degree between reply text and conversation text, thereby assisting the interaction between users and intelligent robots, improving the text matching degree between reply text and conversation text, and helping to improve the accuracy of the first score based on the two. Further, by combining the first score and the reference score to generate evaluation information, the reference score may include the second score of the first user in the second time period, and may also include the third score of the second user, or both. It can be seen that the evaluation information integrates multiple scores, so that the generated evaluation information can comprehensively measure the real situation of the user and improve the accuracy of the evaluation result. In the field of person-job matching, it can not only assist job seekers to efficiently find jobs that match their abilities, but also assist employers to obtain more comprehensive information about job seekers, thereby effectively assisting employers in recruitment. In training fields such as pre-job training and language training, it can assist users to fully understand their real learning situation.

[0030] Figure 1 A schematic diagram of an application scenario of an information processing method provided in an embodiment of the present application. Figure 1 As shown, the application scenario provided by the embodiment of the present application includes a computer device 100, a first terminal device 200 corresponding to a first user, and second terminal devices (eg, Figure 1 The first user and the second user can be a third user. In the field of person-job matching, the first user can be a job seeker, and the second user can be a user with the same interview position as the job seeker; the third user can be an interviewer, for example, the third user can be a virtual character (for example, a cartoon character), and the third user can also be an agent, wherein the agent can be a corporate representative, such as a human resources representative, a corporate department head, and the like, and the agent can also be a third party that provides job information, such as a headhunter. In the field of pre-job training, language training, and other training, the first user and the second user can be learners, for example, the second user's entry position is the same as the first user's entry position, and the second user's learning language is the same as the first user's learning language; the third user can be a job trainer, a language trainer, and the like.

[0031] The first terminal device 200 and the second terminal device 300 can communicate with the computer device 100 and the third terminal device 400 through a network. The third terminal device 400 can also communicate with the computer device 100 through a network. The network can be a wireless network or a wired network. The wireless network includes, but is not limited to: a cellular network, a wireless local area network (Wireless Fidelity, Wi-Fi), an infrared network or a Bluetooth network. The wired network includes, but is not limited to: Ethernet, a universal serial bus (universal serial bus, USB) or a controller local area network. Figure 1 As shown, different first users or second users can initiate the interaction process through different terminal devices. For example, the first user initiates the interaction process through the first terminal device 200, and the second user 1 initiates the interaction process through the second terminal device 300.

[0032] The first terminal device 200, the second terminal device 300 and the third terminal device 400 in the present application can be a personal computer, a tablet computer, a handheld device, a vehicle-mounted device, a wearable device, a computer device or other processing devices connected to a wireless modem, etc. In different networks, the first terminal device 200 or the second terminal device 300 can be called different names, such as: user equipment, access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device, cellular phone, cordless phone, personal digital assistant (PDA), 5G network or terminal equipment in future evolution network, etc. The third terminal device 400 in the present application can also be an intelligent robot.

[0033] The computer device 100 in the present application may be any electronic product that can interact with a user, such as a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), a game console, an interactive network television (IPTV), etc. The computer device 100 may include a network device and / or a user device. The network device includes, but is not limited to, a single network electronic device, an electronic device group consisting of multiple network electronic devices, or a cloud consisting of a large number of hosts or network electronic devices based on cloud computing.

[0034] In another embodiment, the third terminal device 400 can be integrated in the computer device 100 so that the computer device 100 can implement part or all of the functions of the third terminal device 400. For example, the computer device 100 receives a conversation text sent by the first terminal device 200 or the second terminal device 300. After generating the target prompt information, the computer device 100 can generate a reply text of the third terminal device 400 in response to the target prompt information.

[0035] In another embodiment, the first terminal device 200 or the second terminal device 300 can be integrated in the computer device 100 so that the computer device 100 can implement part or all of the functions of the first terminal device 200 or the second terminal device 300. For example, the computer device 100 receives a conversation text sent by the third terminal device 400. After generating the target prompt information, the computer device 100 can generate a reply text of the first terminal device 200 or the second terminal device 300 in response to the target prompt information.

[0036] It should be noted that Figure 1 The number and connection relationship of each device are only for illustration and do not constitute a limitation to the embodiments of the present application.

[0037] like Figure 2 , which is a flow chart of an information processing method provided by an embodiment of the present application. The information processing method is applied to a computer device, for example, Figure 1 The computer device 100. According to different requirements, the order of the steps in the flowchart can be changed, and some steps can be omitted.

[0038] S201, generating target prompt information for the conversation text within a first time period.

[0039] In at least one embodiment of the present application, a computer device may provide a graphical interactive interface as a conversation interface, and receive a conversation text input by a user based on the conversation interface. The conversation may include a text conversation, a voice conversation, or a video conversation, etc. For example, the conversation text may be text information input by the first user or the third user in the conversation interface. The conversation text may also be text converted from the input audio by the first user or the third user during a voice interaction (such as a telephone interaction) or a video interaction. The computer device may display the conversation text converted from the input audio in the conversation interface. For ease of understanding, please refer to Figure 3A , Figure 3A This is a schematic diagram of the conversation text provided in the embodiment of the present application displayed in the dialogue interface. Figure 3A As shown, the first user can enter "I want to find a part-time job" in the dialogue interface.

[0040] The computer device may generate target prompt information for the conversation text. The target prompt information may include the text intent of the conversation text. The text intent may represent the user intent obtained based on the conversation text. The target prompt information may also include relevant user information (e.g., personality traits, income information, experience information, etc.) of the user (e.g., job seeker) who inputs the conversation text. The target prompt information may be used to guide another user (e.g., interviewer) to respond to the conversation text. For ease of understanding, please refer to Figure 3B , Figure 3B is a schematic diagram of displaying target prompt information in a dialogue interface provided by an embodiment of the present application. Figure 3B As shown, the first user can enter the conversation text "I want to find a part-time job" in the dialogue interface. The computer device generates the target prompt information "I (current income is 3,000, introverted personality) want to find a part-time job" based on the conversation text "I want to find a part-time job" entered by the first user in the dialogue interface, and displays the target prompt information in the dialogue interface to assist the third user in understanding the situation of the first user and guide the response to the conversation text.

[0041] In at least one embodiment of the present application, a computer device generates target prompt information for a conversation text within a first time period, including: the computer device identifies a first text intent of the conversation text, extracts key information from the conversation text based on the first text intent through an information extraction model, and generates target prompt information based on the first text intent and the key information.

[0042] Among them, the first text intent can be obtained by identifying the conversation text based on a pre-trained intent recognition model, and the intent recognition model can be trained based on the pre-trained language model of Roberta and others. The training method of the intent recognition model includes: taking the user text (such as the user text of the job applicant, the user text of the interviewer) and the annotated intent corresponding to the user text (for example, the annotation information including the intent label) as training samples, inputting the user text into the pre-trained language model, obtaining the predicted label corresponding to the user text, using the back propagation algorithm and the gradient descent algorithm, updating the pre-trained language model based on the annotated intent and the predicted label until the training stop condition is met. For example, the training stop condition may include but is not limited to the loss value of the pre-trained language model being less than a preset loss threshold, the number of training times of the pre-trained language model being greater than a preset number, etc., and the intent recognition model can be obtained. By identifying the conversation text with the pre-trained intent recognition model to obtain the first text intent, the purpose, intention or intention expressed in the conversation text can be determined, and support can be provided for subsequent tasks such as dialogue management and response generation.

[0043] The information extraction model includes: a sentence coding layer, a phrase relationship coding layer and an entity recognition layer. Among them, the sentence coding layer may include a vocabulary coding table, and the vocabulary coding table includes a mapping relationship between multiple words and codes. The phrase coding layer may include an algorithm framework for identifying phrase relationships, such as a dependency analysis algorithm. Phrase relationships can represent the grammatical relationship of phrases in a text. For example, phrase relationships include, but are not limited to: subject-predicate relationships, verb-object relationships, adverbial-verbal relationships, etc. Phrase relationships are relative, and phrase relationships involve two or more phrases. The "phrase" referred to in this application can be a language unit in any language and divided according to any rules, such as Chinese characters or words, or English words. The phrase coding layer may also include a mapping table for encoding phrase relationships. The entity recognition layer is used to obtain entities contained in the text, and the entity recognition layer can be trained based on training texts with named entity annotation information in historical conversation data. In the process of extracting key information from the conversation text based on the first text intent, the information extraction model can comprehensively consider multiple aspects such as vocabulary, phrase relations, sentence structure, and entity information in the conversation text, and accurately extract key information that is directly related to the first text intent or supports the understanding of the first text intent through a multi-layer encoding and recognition mechanism, thereby more accurately determining the user intent corresponding to the conversation text.

[0044] The target prompt information can be obtained by processing the first text intent and key information based on the pre-trained prompt information generation model. The prompt information generation model includes a template library and an information library. The template library includes the mapping relationship between the role information of the user role and the optimal template. For example, the user role is a job-seeking role, and the template library stores the corresponding relationship between the personality characteristics, income information and the optimal template of multiple users. For another example, the user role is an agent role, and the template library stores the corresponding relationship between the personality characteristics, experience information and the optimal template of multiple users. The information library stores the mapping relationship between any template and intent in the template library and multiple initial prompt information. For the training process of the prompt information generation model, please refer to Figure 5 Updates to the template library shown, and see Figure 6 Updates to the repository shown.

[0045] This embodiment extracts the conversation text in combination with the first text intent, which can improve the accuracy of extracting key information, and further combines the first text intent and key information to generate accurate target prompt information, thereby improving the accuracy of guiding users.

[0046] In at least one embodiment of the present application, a computer device extracts key information from a conversation text based on a first text intent through an information extraction model, including: the computer device uses the first text intent and the conversation text as input text, encodes the input text through a sentence encoding layer, and obtains a first encoding vector. The computer device identifies the phrase relationship in the input text through a phrase encoding layer and encodes the identified phrase relationship to obtain a second encoding vector. The computer device generates encoding information of the input text based on the first encoding vector and the second encoding vector, performs entity recognition on the encoding information based on an entity recognition layer, and obtains key information. Among them, the encoding information can be a concatenated vector of the first encoding vector and the second encoding vector, and the encoding information can also be the sum or product of corresponding encoding elements in the first encoding vector and the second encoding vector.

[0047] Combination Figure 4 The specific structure of the information extraction model in the figure illustrates the extraction of key information. The computer device performs encoding conversion on the input text "I want to find a part-time job" based on the sentence encoding layer to obtain a first encoding vector, and performs encoding conversion on the phrase relationship identified from the input text based on the phrase encoding layer to obtain a second encoding vector. Combining the first encoding vector and the second encoding vector, the encoding information of the input text is generated, and entity recognition is performed on the encoding information based on the entity recognition layer to obtain the key information "looking for a part-time job".

[0048] This embodiment directly encodes the input text through the sentence encoding layer, which can improve the comprehensiveness of the mapping of the first encoding vector to the input text. By encoding the input text through the phrase encoding layer, the phrase relationship of the input text can be encoded from the perspective of grammatical relationship. The encoding information is generated according to the first encoding vector and the second encoding vector, which can improve the mapping accuracy of the encoding information to the input text. Furthermore, based on the entity recognition layer, the key information in the input text can be identified from the encoding information, thereby improving the accuracy of extracting the key information.

[0049] In at least one embodiment of the present application, a computer device generates target prompt information based on a first text intent and key information, including: the computer device obtains a prompt template based on a first user role corresponding to a conversation text, obtains initial prompt information corresponding to the prompt template and the key information, and generates the target prompt information based on an information score of the initial prompt information and the first text intent.

[0050] The first user role may include a job seeker role, an agent role, etc. The first user role may also include a learner role, a trainer role, etc. If the first user role is a job seeker role, the computer device obtains a corresponding template (e.g., a template with the highest matching degree) from the template library as a prompt template based on the personality characteristics of the first user corresponding to the conversation text and the income information of the first user. If the first user role is an agent role, the computer device obtains a corresponding template from the template library as a prompt template based on the personality characteristics of the agent corresponding to the conversation text and the experience information of the agent.

[0051] The process of generating the target prompt information includes: merging the initial prompt information with the highest information score and the first text intention as the target prompt information. For example, the optimal prompt information is: income is X, personality is y, and the first text intention is: I want to find a part-time job. Then the target prompt information can be: I (current income is 3,000, personality is introverted) want to find a part-time job.

[0052] This embodiment can obtain a reasonable prompt template from the template library for application through the first user role, and can obtain initial prompt information suitable for the user through information scoring, thereby improving the rationality of the target prompt information.

[0053] S202, obtaining a reply text in response to the target prompt information.

[0054] In at least one embodiment of the present application, the reply text may be a reply text of another user other than the user corresponding to the conversation text. For example, if the conversation text is an input text of a first user such as a job seeker or a language learner, the reply text may be a reply text of a third user such as an interviewer or a trainer; if the conversation text is an input text of a third user such as an interviewer or a trainer, the reply text may be a reply text of the first user such as a job seeker or a language learner. For ease of understanding, please refer to Figure 3C , Figure 3C is a schematic diagram of displaying reply text in a conversation interface provided by an embodiment of the present application. Figure 3C As shown, the job seeker can enter the conversation text "I want to find a part-time job" in the dialogue interface, and the computer device generates and displays the target prompt information corresponding to the conversation text in the dialogue interface: "I (current income is 3,000, and my personality is introverted) want to find a part-time job." The interviewer can respond to the target prompt information in the dialogue interface and enter the reply text "What is your expected monthly salary" in the dialogue interface.

[0055] In another embodiment, the reply text can be obtained by recognizing the reply audio based on a preset audio recognition model, and the reply audio can be the voice output by the first user or the third user. In different embodiments, the reply text can also be converted from text information to audio for output. In actual applications, there is no restriction on the input method or output method of the reply text.

[0056] In another embodiment, the computer device may generate corresponding target prompt information based on the reply text, and obtain the user's response text corresponding to the conversation text, and the response text is used to respond to the target prompt information corresponding to the reply text. Similarly, there may be multiple rounds of conversations between the first user and the third user.

[0057] S203: Score the conversation text and the reply text to obtain a first score of the first user in a first time period.

[0058] In at least one embodiment of the present application, a preset scoring model is used to score the conversation text and the reply text to obtain a first score of the first user. The preset scoring model includes sub-scoring models corresponding to multiple scoring dimensions, wherein each scoring dimension can correspond to a sub-scoring model. For example, multiple sub-scoring models include, but are not limited to: a scoring model for the relevance of a reply, a scoring model for the understanding of one's own experience, a scoring model for the detail of one's own job-seeking intention, a scoring model for the understanding of the target occupation, a scoring model for fluency of expression, an impression score scoring model, and the like. The training process of each sub-scoring model includes: loading a pre-trained language model, and training the pre-trained language model using the training data to obtain a corresponding sub-scoring model. The training samples of each sub-scoring model include conversation text and reply text, and a score label of the text (such as a score value or a score range). The training process of each sub-scoring model is similar to the training process of the intent recognition model, and this application will not repeat it.

[0059] The first score includes the dimension score of the first user in each dimension of the multiple scoring dimensions, and may also include a comprehensive score of the multiple scoring dimensions. In the field of job matching, the first score may be the interview score of the job seeker in a simulated interview scenario. The simulated interview scenario is a practice method that simulates a real interview environment. The simulated interview scenario is usually used to help job seekers prepare for interviews; the first score may also be the interview score of the job seeker in a real interview scenario. In the fields of pre-job training, language training, etc., the first score may be the test score of the first user in a training test.

[0060] In at least one embodiment of the present application, the computer device scores the conversation text and the reply text based on multiple sub-scoring models to obtain multiple dimension scores. This embodiment scores the conversation text and the reply text through multiple sub-scoring models, and can comprehensively determine the interview score from multiple scoring dimensions to improve the comprehensive evaluation of the first user.

[0061] S204: Determine a reference score, and generate evaluation information for the first user based on the first score and the reference score.

[0062] In at least one embodiment of the present application, the reference score includes at least one of the second score of the first user in the second time period and the third score of the second user. The second score includes the scores of the first user in multiple second time periods, and the multiple second time periods are earlier than the first time period. In the field of person-job matching, the second score may include: the interview scores of the first user in simulated interview scenarios in multiple second time periods. In training fields such as pre-job training and language training, the second score may include: the scores of the first user in training test scenarios in multiple second time periods.

[0063] In the field of job matching, the second user can include all users who have the same job as the first user, and the third score can be regarded as the interview score of the second user, which is used to distinguish it from the interview score of the first user. For example, the third score can include: the interview score of the second user in a simulated interview scenario, and the interview score of the second user in a real interview scenario. In the field of pre-job training, language training, etc., the second user can include all users who have the same job as the first user, and the second user can also include all users who learn the same language as the first user; the third score can include: the score of the second user in the training test scenario.

[0064] In at least one embodiment of the present application, the computer device determines the second user based on the user type of the first user and determines the session time of the session text. The computer device determines a third score based on the second user and the session time.

[0065] The session time may be the input time of the session text or the duration of the session. For example, the session time may be 10:00; the session time may also include a time period determined based on the input time and a preset duration. For example, if the preset duration is 20 minutes and the input time is 10:00, the session time may be 10:00-10:20. The third rating may include the score of the second user corresponding to the session time.

[0066] This embodiment can determine the second user corresponding to the first user through the user type, and then through the session time of the conversation text, can reasonably determine the score corresponding to the second user in the session time as the third score, thereby improving the rationality of the third score.

[0067] In at least one embodiment of the present application, the evaluation information includes first evaluation information, the reference score includes a third score of the second user, and the computer device generates evaluation information for the first user based on the first score and the reference score, specifically including: the computer device calculates a first score change value based on the first score and the third score, and generates the first evaluation information based on the first score change value and a first preset template.

[0068] The first score change value may include the difference between the first score and the third score, and the first score change value may also include the ratio of the first score to the third score, etc. The first preset template is used to indicate the score difference between multiple users. For example, the first preset template is: how many points the first user has higher than the second user.

[0069] This embodiment combines the first score and the third score to determine the score change value between the first user and the second user corresponding to the conversation occurring in the same time period, thereby achieving a horizontal comparative evaluation of the first user and improving the analyzability and content richness of the evaluation information.

[0070] In at least one embodiment of the present application, the evaluation information includes second evaluation information, the reference score includes the second score of the first user, and the computer device generates evaluation information for the first user based on the first score and the reference score, specifically including: the computer device calculates a second score change value based on the first score and the second score, and generates the second evaluation information based on the second score change value and a second preset template.

[0071] The second score change value includes the score change of the first user at multiple times, the second score change value may include the difference between the first score and the second score, the second score change value may also include the ratio of the first score to the second score, etc. The second preset template is used to indicate the score change of the first user at multiple times, for example, the second preset template is: how many points the first user has improved.

[0072] This embodiment can evaluate the stability of the first user by comparing the scores of the same user at different times.

[0073] In the information processing method of this embodiment, by generating target prompt information from the conversation text, the user's reply to the conversation text can be guided, and the text matching degree between the reply text and the conversation text can be improved, thereby assisting the interaction between the user and the intelligent robot, improving the text matching degree between the reply text and the conversation text, and helping to improve the accuracy of the first score based on the two. Further, by combining the first score and the reference score to generate evaluation information, the reference score may include the second score of the first user in the second time period, and may also include the third score of the second user, or both. It can be seen that the evaluation information integrates multiple scores, so that the generated evaluation information can comprehensively measure the real situation of the user and improve the accuracy of the evaluation results.

[0074] like Figure 5 As shown in FIG. 1 , it is a flow chart of updating the template library provided by an embodiment of the present application. Figure 5 As shown, it specifically includes steps S501-S503:

[0075] S501: Obtain a prediction template based on a second user role corresponding to session data in a third time period.

[0076] In at least one embodiment of the present application, the third time period is earlier than the first time period, and the conversation data in the third time period may include the communication text generated by the third user and the first user during the communication process. The second user role may include a job seeker role, an agent role, a learner role, a trainer role, etc.

[0077] In some embodiments of the present application, an initial template library can be created in advance, the initial template library includes multiple setting templates, and a mapping relationship between the role information of the user role and the multiple setting templates is established, wherein the setting template can be extracted from the big data conversation text, and the setting template can also be customized by the development user. The prediction template is a template in the template library that successfully matches the role information of the second user role. The role information of the job-seeking role may include but is not limited to a combination of one or more information: personality traits, income information, and job positions. The role information of the agent role may include but is not limited to a combination of one or more information: personality traits, experience information. For example, the data stored in the initial template library is as shown in Table 1:

[0078] Table 1

[0079] User roles Personality and characteristics Income Information Experience information Setting template Job role Extroversion 1000 NTD Template 1, Template 2, Template 3 Agent Role Introversion 5 years Template 4, Template 5

[0080] S502: Select a target prediction template from the prediction templates based on the first annotation score of each prediction template.

[0081] In at least one embodiment of the present application, the first annotation score may be a score of the prediction template by multiple scoring users corresponding to the second user role. For example, the second user role is a job seeker role, and the first annotation score may be a score of the prediction template by multiple job seekers. The target prediction template may be a prediction template with the highest comprehensive score, wherein the calculation process of the comprehensive score of each prediction template includes: calculating the average of multiple first annotation scores as the comprehensive score of each prediction template. In another embodiment, the calculation process of the comprehensive score of each prediction template also includes: calculating the weighted sum of multiple first annotation scores as the comprehensive score of each prediction template, wherein the weight corresponding to each first annotation score can be determined based on the credit value of the scoring user corresponding to the first annotation score. In the field of person-job matching, the credit value can be determined based on the number of interviews of the scoring user, and the credit value is positively correlated with the number of interviews. In the training fields such as pre-job training and language training, the credit value can be determined based on the number of tests of the scoring user, and the credit value is positively correlated with the number of tests.

[0082] S503: Update the template library based on the second user role and the target prediction template.

[0083] In at least one embodiment of the present application, updating the template library includes: the computer device replaces multiple prediction templates in the template library that successfully match the second user role based on the target prediction template.

[0084] like Figure 6 As shown in FIG. 1 , it is a flowchart of updating the information base provided by an embodiment of the present application. Figure 6 As shown, it specifically includes steps S601-S603:

[0085] S601, based on the target prediction template in the template library and the second text intention of the conversation data in the third time period, obtain a prediction prompt corresponding to the target prediction template from the information library.

[0086] In at least one embodiment of the present application, the target prediction template may be a template in the template library that successfully matches the role information of the second user role, and the second user role is the user role corresponding to the session data in the third time period. The second text intent may be obtained by identifying the session data in the third time period based on the intent recognition model. For an introduction to the intent recognition model, please refer to Figure 2 Related content.

[0087] In at least one embodiment of the present application, an initial information library may be pre-created, the initial information library including a plurality of setting prompts, and a mapping relationship between any template in the template library, a setting intent, and a plurality of setting prompts is established. The plurality of prediction prompts include setting prompts in the information library that successfully match both the target prediction template and the second text intent.

[0088] S602: For each prediction prompt, based on the second annotation score of the prediction prompt, calculate a third annotation score of the prediction prompt.

[0089] In at least one embodiment of the present application, the second annotation score can be the score of the prediction prompt by multiple scoring users corresponding to the second user role. For example, if the second user role is a job seeker role, the second annotation score can be the score of the prediction prompt by multiple job seekers. The calculation process of the third annotation score includes: calculating the mean of multiple second annotation scores as the third annotation score of each prediction prompt. In another embodiment, the calculation process of the third annotation score also includes: calculating the weighted sum of multiple second annotation scores as the third annotation score of each prediction prompt, wherein the weight corresponding to each second annotation score can be determined based on the credit value of the scoring user corresponding to the second annotation score. In the field of person-job matching, the credit value can be determined based on the number of interviews of the scoring user, and the credit value is positively correlated with the number of interviews. In training fields such as pre-job training and language training, the credit value can be determined based on the number of tests of the scoring user, and the credit value is positively correlated with the number of tests.

[0090] S603: Update the information base based on the target prediction template, the prediction prompt and the third annotation score of each prediction prompt.

[0091] In at least one embodiment of the present application, the updating process of the information library includes: the computer device updates the initial scores of multiple prediction prompts corresponding to the target prediction template in the information library based on the third annotation score of each prediction prompt.

[0092] This embodiment combines the first annotation score of each prediction template by the scoring user to reasonably determine the target prediction template from multiple prediction templates, obtains multiple prediction prompts from the information library for scoring analysis based on the target prediction template, and can reasonably determine the third annotation score of each prediction prompt, and then updates the template library based on the target prediction template, and updates the score of the corresponding prediction prompt in the information library based on the third annotation score of each prediction prompt, so that the target prompt information can be accurately extracted in combination with the template library and the information library in the future.

[0093] like Figure 7 As shown in FIG. 1 , it is a flowchart of calculating the interview score provided by an embodiment of the present application. Figure 7 As shown, it specifically includes steps S701-S703:

[0094] S701, taking the conversation text and the reply text as the final text, and identifying the final text based on a preset skill recognition model to determine the user skills of the first user.

[0095] In at least one embodiment of the present application, the training process of the preset skill recognition model includes: performing sequence annotation on the conversation text in the fourth time period, obtaining the skill point name and the position information of the skill point name in the conversation text in the fourth time period, training the pre-trained language model based on the conversation text in the fourth time period, the skill point name and the corresponding position information, until the skill recognition accuracy of the pre-trained language model for the conversation text in the fourth time period after training is greater than the preset accuracy threshold, and obtaining the preset skill recognition model. Among them, the fourth time period is earlier than the first time period, and the pre-trained language model can be a Roberta model.

[0096] Among them, the skill recognition accuracy can be determined based on the predicted skills and skill point names of the conversation text in the fourth time period by the pre-trained language model. The calculation process of the skill recognition accuracy may include: counting the first number of predicted skills with the same skill point name, determining the larger value of the total number of skill point names and the total number of predicted skills as the second number, and determining the skill recognition accuracy based on the ratio of the first number to the second number. For example, there are 10 skill point names, namely skills 1 to skill 10, and there are 8 predicted skills, namely skills 1 to skill 8. After determination, the first number is 8 and the second number is 10, then the skill recognition accuracy can be 8 / 10=0.8. For another example, there are 10 skill point names, namely skills 1 to skill 10, and there are 11 predicted skills, namely skills 1 to skill 11. After determination, the first number is 10 and the second number is 11, then the skill recognition accuracy can be 10 / 11.

[0097] For example, the interviewer's conversation text in the final text may be: "Which libraries in programming language AAA are you best at?"; the job seeker's response text in the final text may be: "The libraries I am most familiar with are 001 library, 002 library, ... and so on."; the job seeker's user skills identified by the preset skill recognition model may include: 001 library, 002 library, ... and so on.

[0098] S702, calculating the text similarity between the final text and the preset standard text corresponding to the user skill.

[0099] In at least one embodiment of the present application, the preset standard text may be a model text corresponding to the user's skills, for example, the preset standard text may be: "I master programming of library 001". The text similarity may be calculated based on the cosine distance formula between the representation vector of the final text and the representation vector of the preset standard text.

[0100] S703: Calculate a first score of the first user in a first time period based on the text similarity.

[0101] In at least one embodiment of the present application, the calculation process of the first score includes: weighting and calculating the text similarity based on the skill weight corresponding to the user skill to obtain the first score. For example, for the position of JAVA engineer, if the user skills of the job seeker include Java language and 001 library programming, the weight corresponding to "Java language" is 0.3, and the weight corresponding to "001 library programming" is 0.1, if the text similarity between the final text and the preset standard text corresponding to "Java language" is 2, and the text similarity between the final text and the preset standard text corresponding to "001 library programming" is 1, then the first score is 0.3*2+0.1*1=0.7.

[0102] This embodiment recognizes the final text through a preset skill recognition model, which can quickly identify the user's skills from the final text and improve the calculation efficiency of the interview score.

[0103] like Figure 8 As shown, it is a flowchart of another information processing method provided by an embodiment of the present application. Figure 8 As shown, steps S801-S808 are included:

[0104] S801: Acquire session data within a third time period from a session library.

[0105] In at least one embodiment of the present application, the session library stores session data between the third user and the first user in an interaction scenario.

[0106] S802: Update the template library and the information library based on the session data in the third time period.

[0107] In at least one embodiment of the present application, the computer device updates the templates corresponding to different roles in the template library and the prompt information corresponding to different roles in the information library respectively according to the session data in the third time period corresponding to the different roles.

[0108] For details of step S802, please refer to the above Figure 5 and Figure 6 The detailed description is not repeated here.

[0109] S803, in the simulation test, generating target prompt information for the conversation text in the first time period.

[0110] In the field of person-job matching, a mock test can be a simulated interview scenario. In training fields such as pre-job training and language training, a mock test can be a simulated test scenario.

[0111] S804, obtaining a reply text in response to the target prompt information.

[0112] In at least one embodiment of the present application, the reply text may be any text in the limited answer corresponding to the target prompt information. After executing step S804, the computer device may use the reply text as the conversation text in step S803, and repeatedly execute steps S803-S804 until the simulation test is completed.

[0113] S805, after the simulation test is finished, scoring is performed based on the conversation text and the reply text to obtain a first score of the first user in the first time period.

[0114] S806: Determine a reference score, and generate evaluation information for the first user based on the first score and the reference score.

[0115] For details of steps S803-S806, please refer to the above Figure 2 and Figure 7 The detailed description is not repeated here.

[0116] S807, in a real test scenario, execute steps S801 to S806.

[0117] In the field of person-job matching, the real test can be a real interview scenario. In training fields such as pre-job training and language training, the real test can be a real test scenario.

[0118] S808, storing the conversation text and reply text of the actual test scenario into the conversation library.

[0119] This embodiment stores the conversation text and reply text of the real test scenario in the conversation library. Since the conversation text and reply text of the real test scenario are closer to reality, the template library and the information library are updated through the conversation data in the conversation library, thereby improving the update effect of the template library and the information library.

[0120] like Fig. 9 , is a functional module diagram of an information processing device provided by an embodiment of the present application. The information processing device 11 includes a generating unit 110, an acquiring unit 111, a scoring unit 112, a selecting unit 113, an updating unit 114 and a calculating unit 115. The module / unit referred to in the present application refers to a type of information processing device that can be processed by a processor (e.g. Fig.10 A series of computer readable instruction segments that can be obtained by the processor 1101 shown in the figure and can perform fixed functions and are stored in a memory (eg Fig.10 In the memory 1102 shown.

[0121] The generation unit 110 is used to generate target prompt information for the conversation text within the first time period; the acquisition unit 111 is used to acquire the reply text in response to the target prompt information; the scoring unit 112 is used to score based on the conversation text and the reply text to obtain the first score of the first user in the first time period; the generation unit 110 is also used to determine a reference score, and generate evaluation information for the first user based on the first score and the reference score, the reference score including at least one of the second score of the first user in the second time period and the third score of the second user.

[0122] In one embodiment, the generation unit 110 is specifically used to: identify a first text intent of a conversation text; based on the first text intent, extract key information from the conversation text through an information extraction model; based on the first text intent and the key information, generate target prompt information.

[0123] In one embodiment, the information extraction model includes: a sentence coding layer, a phrase relationship coding layer and an entity recognition layer, and a generation unit 110 is specifically used to: take the first text intention and the conversation text as input text, encode the input text through the sentence coding layer, and obtain a first coding vector; identify the phrase relationship in the input text through the phrase coding layer and encode the identified phrase relationship to obtain a second coding vector; generate coding information of the input text based on the first coding vector and the second coding vector; perform entity recognition on the coded information based on the entity recognition layer to obtain key information.

[0124] In one embodiment, the generation unit 110 is further specifically used to: obtain a prompt template based on the first user role corresponding to the conversation text; obtain initial prompt information corresponding to the prompt template and key information; and generate target prompt information based on the information score of the initial prompt information and the first text intention.

[0125] In one embodiment, the prompt template is obtained from a template library, and the acquisition unit 111 is also used to obtain a prediction template based on the second user role corresponding to the session data in a third time period; the selection unit 113 is used to select a target prediction template from the prediction template based on the first annotation score of each prediction template; the updating unit 114 is used to update the template library based on the second user role and the target prediction template.

[0126] In one embodiment, the initial prompt information is obtained from an information base, and the acquisition unit 111 is further used to obtain a prediction prompt corresponding to a target prediction template from the information base based on the target prediction template and the second text intent of the conversation data; the calculation unit 115 is used to calculate a third annotation score of the prediction prompt based on the second annotation score of the prediction prompt for each prediction prompt; the updating unit 114 is further used to update the information base based on the target prediction template, the prediction prompt and the third annotation score of each prediction prompt.

[0127] In one embodiment, the scoring unit 112 is specifically used to: take the conversation text and the reply text as the final text, and identify the final text based on a preset skill recognition model to determine the user skills of the first user; calculate the text similarity between the final text and the preset standard text corresponding to the user skills; and calculate the first score of the first user in the first time period based on the text similarity.

[0128] In one embodiment, the generating unit 110 is further specifically configured to: determine the second user based on the user type of the first user, and determine the session time of the session text; and determine the third score based on the second user and the session time.

[0129] In one embodiment, the generating unit 110 is further specifically configured to: calculate a score change value according to the first score and the reference score; and generate evaluation information according to the score change value and a preset template.

[0130] In multiple embodiments of the present application, by generating target prompt information from the conversation text, it is possible to guide the user's reply to the conversation text, improve the text matching degree between the reply text and the conversation text, thereby assisting the user's interaction with the intelligent robot, improving the text matching degree between the reply text and the conversation text, and helping to improve the accuracy of the first score based on the two. Further, by combining the first score and the reference score to generate evaluation information, the reference score may include the second score of the first user in the second time period, and may also include the third score of the second user, or both. It can be seen that multiple scores are integrated in the evaluation information, so that the generated evaluation information can comprehensively measure the user's true situation and improve the accuracy of the evaluation results.

[0131] Fig.10 is a schematic diagram of the structure of a computer device for implementing an information processing method provided in an embodiment of the present application. Optionally, Fig.10 The computer device 100 is shown for executing Figure 2-Figure 8 The method shown.

[0132] The computer device 100 includes at least one processor 1101 , a memory 1102 , and at least one network interface 1103 .

[0133] The processor 1101 is, for example, a general-purpose central processing unit (CPU), a network processor (NP), a graphics processing unit (GPU), a neural-network processing unit (NPU), a data processing unit (DPU), a microprocessor, or one or more integrated circuits for implementing the scheme of the present application. For example, the processor 1101 includes an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The PLD is, for example, a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0134] The memory 1102 is, for example, a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. Optionally, the memory 1102 exists independently and is connected to the processor 1101 via an internal connection 1104. Alternatively, the memory 1102 and the processor 1101 are optionally integrated together.

[0135] The network interface 1103 uses any transceiver-like device for communicating with other devices or communication networks. The network interface 1103 includes, for example, at least one of a wired network interface or a wireless network interface. The wired network interface is, for example, an Ethernet interface. The Ethernet interface is, for example, an optical interface, an electrical interface, or a combination thereof. The wireless network interface is, for example, a wireless local area network (WLAN) interface, a cellular network interface, or a combination thereof.

[0136] In some embodiments, the processor 1101 includes one or more CPUs, such as Fig.10 CPU0 and CPU1 are shown in the figure.

[0137] In some embodiments, the computer device 100 optionally includes multiple processors, such as Fig.10 1 and 1105. Each of these processors is, for example, a single-core processor (single-CPU), or a multi-core processor (multi-CPU). The processor here optionally refers to one or more devices, circuits, and / or processing cores for processing data (such as computer program instructions).

[0138] In some embodiments, the computer device 100 further includes an internal connection 1104. The processor 1101, the memory 1102, and the at least one network interface 1103 are connected via the internal connection 1104. The internal connection 1104 includes a path to transmit information between the above components. Optionally, the internal connection 1104 is a single board or a bus. Optionally, the internal connection 1104 is divided into an address bus, a data bus, a control bus, etc.

[0139] In some embodiments, the computer device 100 further includes an input-output interface 1106. The input-output interface 1106 is connected to the internal connection 1104.

[0140] Optionally, the processor 1101 implements the method in the above embodiment by reading the program code 910 stored in the memory 1102, or the processor 1101 implements the method in the above embodiment by the program code stored internally. In the case where the processor 1101 implements the method in the above embodiment by reading the program code 910 stored in the memory 1102, the memory 1102 stores the program code for implementing the method provided in the embodiment of the present application.

[0141] For more details on how the processor 1101 implements the above functions, please refer to the descriptions in the previous method embodiments, which will not be repeated here.

[0142] This embodiment also provides a computer storage medium, in which computer instructions are stored. When the computer instructions are executed on a computer device, the computer device executes the above-mentioned related method steps to implement the information processing method in the above-mentioned embodiment.

[0143] This embodiment further provides a computer program product. When the computer program product is run on a computer device, the computer device is caused to execute the above-mentioned related steps to implement the information processing method in the above-mentioned embodiment.

[0144] In addition, an embodiment of the present application also provides a device, which can specifically be a chip, component or module, and the device may include a connected processor and memory; wherein the memory is used to store computer execution instructions, and when the device is running, the processor can execute the computer execution instructions stored in the memory so that the chip executes the information processing method in the above-mentioned method embodiments.

[0145] Among them, the computer device, computer storage medium, computer program product or chip provided in this embodiment is used to execute the corresponding method provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method provided above, and will not be repeated here.

[0146] Through the description of the above implementation methods, technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0147] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0148] The unit described as a separate component may or may not be physically separated, and the component shown as a unit may be one physical unit or multiple physical units, that is, it may be located in one place or distributed in multiple different places. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiment.

[0149] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0150] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application is essentially 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 storage medium, including several instructions to enable a device (which can be a single-chip microcomputer, chip, etc.) or a processor (processor) to perform all or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store program code.

[0151] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto, and any changes or substitutions within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An information processing method, characterized in that: The method comprises: generating target prompt information for the conversation text in the first time period; Obtaining a reply text in response to the target prompt information; Scoring the conversation text and the reply text to obtain a first score of the first user in the first time period; A reference score is determined, and evaluation information for the first user is generated based on the first score and the reference score, wherein the reference score includes at least one of a second score of the first user in a second time period and a third score of the second user.

2. The information processing method according to claim 1, characterized in that: The step of generating target prompt information for the conversation text within the first time period includes: Identifying a first textual intent of the conversation text; Based on the first text intention, extract key information from the conversation text through an information extraction model; The target prompt information is generated based on the first text intention and the key information.

3. The information processing method according to claim 2, characterized in that: The information extraction model includes: a sentence encoding layer, a phrase relationship encoding layer and an entity recognition layer. The information extraction model is used to extract key information from the conversation text based on the first text intention, including: Taking the first text intention and the conversation text as input text, encoding the input text through the sentence encoding layer to obtain a first encoding vector; Identifying a phrase relationship in the input text through the phrase encoding layer and encoding the identified phrase relationship to obtain a second encoding vector; Generate encoding information of the input text according to the first encoding vector and the second encoding vector; Entity recognition is performed on the encoded information based on the entity recognition layer to obtain the key information.

4. The information processing method according to claim 2, characterized in that: The generating the target prompt information based on the first text intention and the key information includes: Based on the first user role corresponding to the conversation text, obtaining a prompt template; Acquire initial prompt information corresponding to the prompt template and the key information; The target prompt information is generated based on the information score of the initial prompt information and the first text intention.

5. The information processing method according to claim 4, characterized in that: The prompt template is obtained from a template library, and the method further includes: Obtaining a prediction template based on the second user role corresponding to the session data in the third time period; Selecting a target prediction template from the prediction templates based on the first annotation score of each prediction template; Based on the second user role and the target prediction template, the template library is updated.

6. The information processing method according to claim 5, characterized in that: The initial prompt information is obtained from an information base, and the method further includes: Based on the target prediction template and the second text intent of the conversation data, obtaining a prediction prompt corresponding to the target prediction template from an information library; For each prediction prompt, calculating a third annotation score of the prediction prompt based on the second annotation score of the prediction prompt; The information library is updated based on the target prediction template, the prediction prompts and a third annotation score of each prediction prompt.

7. The information processing method according to claim 1, characterized in that: The scoring based on the conversation text and the reply text to obtain a first score of the first user in the first time period includes: Taking the conversation text and the reply text as final text, and identifying the final text based on a preset skill recognition model to determine the user skills of the first user; Calculating the text similarity between the final text and a preset standard text corresponding to the user skill; Based on the text similarity, a first score of the first user in the first time period is calculated.

8. The information processing method according to claim 1, characterized in that: The reference score includes the third score of the second user, and determining the reference score includes: determining the second user based on the user type of the first user, and determining the session time of the session text; The third score is determined based on the second user and the session time.

9. The information processing method according to claim 1, characterized in that: The generating evaluation information of the first user according to the first score and the reference score includes: Calculating a score change value according to the first score and the reference score; The evaluation information is generated according to the score change value and a preset template.

10. An information processing device, characterized in that: The device comprises: A generating unit, used to generate target prompt information for the conversation text within the first time period; An acquisition unit, used for acquiring a reply text in response to the target prompt information; A scoring unit, configured to score the conversation text and the reply text to obtain a first score of the first user in a first time period; The generating unit is further configured to determine a reference score, and generate evaluation information for the first user based on the first score and the reference score, wherein the reference score includes at least one of a second score of the first user in a second time period and a third score of the second user.

11. A computer device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the information processing method according to any one of claims 1 to 9 when executing the computer program.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the information processing method according to any one of claims 1 to 9 is implemented.

13. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the information processing method according to any one of claims 1 to 9 is implemented.

Images