Method, device, medium and electronic equipment for information interaction based on meeting schedule

Abstract

The application relates to a conference schedule-based information interaction method and device, a medium and an electronic equipment, and relates to the technical field of data processing. The method comprises the following steps: inputting first query text information into a preset general large model to determine target user query demand information for a conference schedule; if the target conference theme is not contained in each target schedule information entity, the associated schedule information entity is queried from a conference schedule knowledge graph; first reply text information corresponding to the first query text information is generated and sent to the terminal of the target user; if the target conference theme is contained in each target schedule information entity, after the rationality of the target conference theme is verified, second reply text information corresponding to the first query text information is generated and sent to the terminal of the target user. The application has the effect of improving the matching of the queried conference schedule information and the participant demand.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a conference schedule-based information interaction method and device, medium and electronic equipment. BACKGROUND

[0002] A conference schedule refers to a plan for implementing activities (including ceremonial and auxiliary activities) during a conference according to a unit time. Conference schedule information is a combination system of "time + activity", including conference start / end time, link start / end time, link content, duration, location, responsible person, conference presenter and other elements. In addition, conference schedule information can help participants coordinate action rhythm, for example, participants can accurately lock core content and avoid blind rush according to conference schedule information, and participants can also better coordinate participation and external work according to conference schedule information to avoid delay on both ends. It can be seen that the acquisition of conference schedule information is particularly important for participants, and information interaction in the process of acquiring conference schedule information is also crucial.

[0003] Currently, the information interaction based on the conference schedule usually adopts the following way: participants often input the keywords for searching in the special website of the relevant conference, so as to obtain the conference schedule information that they want to know, but in this way, the special website usually only supports accurate keyword matching, once the participants have fuzzy query behavior or input natural language expression, the information retrieved by the special website may deviate from what the participants want to query, resulting in poor matching of the queried conference schedule information and the needs of the participants. SUMMARY

[0004] In order to improve the matching of the queried conference schedule information and the needs of the participants, the present application provides a conference schedule-based information interaction method, device, medium and electronic equipment.

[0005] In a first aspect of the present application, a conference schedule-based information interaction method is provided, which specifically includes:

[0006] Obtaining first query text information of a target user for a conference schedule, and obtaining a conference schedule knowledge graph, the conference schedule knowledge graph including schedule information entities related to different to-be-held conferences and relationships between different schedule information entities, the schedule information entity being a single schedule information related to the to-be-held conference;

[0007] Inputting the first query text information into a preset general large model to determine query demand information of the target user for the conference schedule, the query demand information including at least one target schedule information entity and a target query intent;

[0008] if each of the target schedule information entities does not contain the target meeting theme, then according to each of the target schedule information entities, an associated schedule information entity is queried from the meeting schedule knowledge graph, the associated schedule information entity and each of the target schedule information entities belong to the same schedule information related to a to-be-held meeting;

[0009] According to the associated schedule information entity and the target query intention, the first query text information corresponding to the first reply text information is generated, and the first reply text information is sent to the terminal of the target user;

[0010] If each of the target schedule information entities contains the target meeting theme, the target meeting theme is reasonably verified, and after the verification passes, the second reply text information corresponding to the first query text information is generated according to the target meeting theme and the target query intention, and the second reply text information is sent to the terminal of the target user.

[0011] By adopting the above technical solution, after obtaining the first query text information and the meeting schedule knowledge graph, the general large model is used to analyze and determine the query demand of the target user for the to-be-held meeting based on the first query text information. Then, when each target schedule information entity does not contain the target meeting theme, the associated schedule information entity associated with the target schedule information entity is quickly queried through the relationship between the entities in the meeting schedule knowledge graph, and is used as the reference content for generating the reply corresponding to the first query text information. Then, the first reply text information that meets the demand of the target user is generated. When each target schedule information entity contains the target meeting theme, in order to more accurately determine the to-be-held meeting for which the target user wants to query the schedule information, the target meeting theme is reasonably verified, and after the verification passes, the second reply text information related to the to-be-held meeting of the target meeting theme is generated in combination with the query intention (target query intention) of the target user. Thus, the matching of the queried meeting schedule information and the demand of the target user is improved.

[0012] In an embodiment, the reasonable verification of the target meeting theme specifically includes:

[0013] Obtain the theme label words of a plurality of meetings for which the historical user has queried schedule information, and determine at least one target label word according to each of the theme label words, the target label word being a theme label word that the historical user pays attention to and is easy to involve in the meeting, and the historical user being consistent with the user portrait of the target user;

[0014] The input historical keywords are obtained when the historical user queries the schedule information of the target meeting, and at least one target keyword is determined according to each of the historical keywords, the target keyword being a historical keyword that is easy for the historical user to input, and the target meeting being a meeting involving a single target label keyword;

[0015] A first weight value of each of the target label keywords is determined, and a second weight value of a target keyword corresponding to each of the target label keywords is determined, the first weight value representing a possibility size of a meeting that the historical user pays attention to involving the target label keyword, and the second weight value representing a possibility size of the historical user inputting the target keyword when querying the schedule information of the target meeting;

[0016] At least one actual label keyword involved in a meeting corresponding to the target meeting theme is obtained, and at least one actual keyword is extracted from the first query text information, and the target meeting theme is reasonably verified according to the actual label keyword, the actual keyword, the first weight value and the second weight value.

[0017] In an embodiment, the reasonably verifying the target meeting theme according to the actual label keyword, the actual keyword, the first weight value and the second weight value specifically includes:

[0018] An actual label keyword consistent with the target label keyword is determined as a key label keyword, and at least one actual keyword existing in each target keyword corresponding to the key label keyword is determined as a key keyword;

[0019] A first product of a first weight value of at least one key label keyword and a second weight value of each key keyword is calculated, and each first product is summed to obtain a first comprehensive result;

[0020] The first comprehensive result is compared with a preset first threshold value, and if the first comprehensive result is greater than the first threshold value, it is determined that the reasonable verification of the target meeting theme is passed;

[0021] If the first comprehensive result is not greater than the first threshold value, it is determined that the reasonable verification of the target meeting theme is not passed.

[0022] In an embodiment, the method further includes:

[0023] When the verification is not passed, if the actual keyword exists in each target keyword corresponding to the target label keyword, the corresponding target label keyword is determined as an important label keyword, and at least one actual keyword existing in each target keyword corresponding to the important label keyword is determined as an important keyword;

[0024] calculating a second product of a first weight value of at least one of the important label words and a second weight value of each of the important keywords, summing each of the second products to obtain a second comprehensive result of the corresponding important label word;

[0025] if the second comprehensive result is greater than a preset second threshold value, determining the corresponding important label word as a to-be-selected label word, sending each of the to-be-selected label words to the terminal, and receiving at least one selected label word sent by the terminal;

[0026] counting a number of the selected label words involved in each of the to-be-held meetings, and adjusting and optimizing the target meeting theme according to the number and the selected label words involved to obtain a final meeting theme;

[0027] generating a third reply text information corresponding to the first query text information according to the final meeting theme and the target query intention, and sending the third reply text information to the terminal of the target user.

[0028] In an implementation manner, the adjusting and optimizing the target meeting theme according to the number and the selected label words involved to obtain a final meeting theme specifically includes:

[0029] summing the second comprehensive results corresponding to the selected label words involved to obtain a third comprehensive result;

[0030] determining a result correction coefficient according to the number, the larger the number is, the larger the result correction coefficient is, and the result correction coefficient is a positive number not less than 1;

[0031] multiplying the result correction coefficient and the third comprehensive result to obtain a corrected result, and if the corrected result is greater than a preset third threshold value, determining a meeting theme of the to-be-held meeting corresponding to the corrected result as the final meeting theme.

[0032] In an implementation manner, the method further includes:

[0033] after the verification passes, if it is detected that the target user is in an inputting state, determining an actual label word consistent with the target label word as a key label word, and calculating a third product of a first weight value of at least one of the key label words and a second weight value of each target keyword corresponding to the key label word;

[0034] summing each third product corresponding to the same target keyword to obtain a target comprehensive result, and if the target comprehensive result is greater than a preset fourth threshold value, determining the corresponding target keyword as a to-be-selected keyword;

[0035] The sub-keyword input by the target user currently is obtained, and a candidate keyword containing the sub-keyword is displayed in the form of prompt text in an input box in the terminal.

[0036] In an implementation, the method further includes:

[0037] If the first query text information of the target user for the conference schedule is not obtained after a preset time period, the current location of the target user, a meeting departure time, and locations of the conference venues of the to-be-held conferences are obtained.

[0038] According to the current location and the locations of the conference venues, an estimated time length for the target user to reach the conference venues is determined, and according to the meeting departure time and the estimated time length, an estimated arrival time of the target user to reach each conference venue of the to-be-held conferences is determined.

[0039] If a conference start time corresponding to the to-be-held conference is later than a corresponding estimated arrival time, the to-be-held conference is determined as a to-be-recommended conference.

[0040] A first weight value of a target label word involved in the to-be-recommended conference is summed to obtain a weight value sum, and if the weight value sum is greater than a preset weight value threshold, schedule information of the to-be-recommended conference is sent to the terminal.

[0041] In a second aspect of the present application, an information interaction device based on a conference schedule is provided, specifically including:

[0042] An information acquisition module is configured to acquire first query text information of a target user for a conference schedule, and acquire a conference schedule knowledge graph, the conference schedule knowledge graph including different conference information entities related to to-be-held conferences and relationships between different conference information entities, and the conference information entity being a single conference information related to the to-be-held conference;

[0043] A demand determination module is configured to input the first query text information into a preset general model to determine query demand information of the target user for the conference schedule, the query demand information including at least one target conference information entity and a target query intent;

[0044] An entity query module is configured to, if the target conference theme is not included in each target conference information entity, query associated conference information entities from the conference schedule knowledge graph according to each target conference information entity, the associated conference information entities belonging to the same conference information related to the to-be-held conference as each target conference information entity;

[0045] The first reply module is configured to generate first reply text information corresponding to the first query text information according to the associated schedule information entity and the target query intention, and send the first reply text information to the terminal of the target user.

[0046] The second reply module is configured to, if the target conference theme is contained in each of the target schedule information entities, perform rationality verification on the target conference theme, and after the verification is passed, generate second reply text information corresponding to the first query text information according to the target conference theme and the target query intention, and send the second reply text information to the terminal of the target user.

[0047] By using the above technical solutions, the information acquisition module acquires the first query text information and the conference schedule knowledge graph, the demand determination module determines the query demand information of the target user for the conference schedule, then the entity query module queries the associated schedule information entity from the conference schedule knowledge graph, the first reply module generates the first reply text information and sends the first reply text information to the terminal of the target user, and finally, the second reply module generates the second reply text information and sends the second reply text information to the terminal of the target user.

[0048] In a third aspect of the present application, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program. When the computer program is loaded and executed by a processor, the method steps of any one of the first aspect are executed.

[0049] In a fourth aspect of the present application, an electronic device is provided, and specifically includes:

[0050] The processor, the memory, and the computer program stored in the memory and capable of running on the processor are configured to load and execute the computer program stored in the memory, so that the electronic device executes the method of any one of the first aspect.

[0051] In summary, the present application includes at least one of the following beneficial technical effects: by the general large model based on the first query text information, the query demand of the target user for the to-be-held meeting is analyzed and determined, and then when the target meeting theme is not included in each target schedule information entity, the associated schedule information entity associated with the target schedule information entity is quickly queried through the relationship between the entities in the meeting schedule knowledge graph, and is used as the reference content for generating the first query text information corresponding to the reply, and then the first reply text information that meets the demand of the target user is generated; when the target meeting theme is included in each target schedule information entity, in order to more accurately determine the to-be-held meeting of the target user who wants to query the schedule information, the rationality of the target meeting theme is verified, and after the verification is passed, the second reply text information related to the to-be-held meeting of the target meeting theme is generated in combination with the query intention of the target user (target query intention), so that the matching of the queried meeting schedule information and the demand of the target user is improved. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 is a flow diagram of a meeting schedule-based information interaction method provided by an embodiment of the present application;

[0053] Figure 2 is a scene architecture diagram of a meeting schedule-based information interaction method provided by an embodiment of the present application;

[0054] Figure 3 is a diagram reflecting the association between the target label word and the target keyword provided by an embodiment of the present application;

[0055] Figure 4 is a structural diagram of a meeting schedule-based information interaction device provided by an embodiment of the present application;

[0056] Figure 5 is a structural diagram of another meeting schedule-based information interaction device provided by an embodiment of the present application.

[0057] Reference signs: 11, information acquisition module; 12, demand determination module; 13, entity query module; 14, first reply module; 15, second reply module; 16, third reply module; 17, input assistance module; 18, meeting recommendation module. DETAILED DESCRIPTION

[0058] In order to enable personnel in the technical field to better understand the technical solutions in the present specification, the technical solutions in the present specification will be clearly and completely described below in combination with the drawings in the embodiments of the present specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments.

[0059] In the description of the embodiments of the present application, the words "exemplarily", "for example", "for instance" or the like are used to represent as an example, illustration or description. Any embodiment or design scheme described as "exemplarily", "for example" or "for instance" in the embodiments of the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. In fact, the words "exemplarily", "for example" or "for instance" are intended to present the relevant concept in a specific manner.

[0060] In the description of the embodiments of the present application, the term "and / or" is only used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent three cases of existence of A alone, existence of B alone and existence of A and B at the same time. In addition, unless otherwise specified, the term "multiple" means two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are only used for description purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Therefore, the features defined with "first" and "second" can explicitly or implicitly include one or more features. The terms "include", "contain", "have" and their variants mean "include but are not limited to", unless otherwise specifically emphasized.

[0061] Referring to Figure 1 The embodiments of the present application disclose a flowchart of a conference schedule-based information interaction method, which can be implemented by relying on a computer program and can also run on a conference schedule-based information interaction device based on the Von Neumann system. The computer program can be integrated in an application or can run as an independent tool class application. Specifically, the computer program comprises:

[0062] S101: Obtain first query text information of a target user for a conference schedule, and obtain a conference schedule knowledge graph.

[0063] Specifically, in the embodiments of the present application, the target user is a user who needs to know the schedule information related to the to-be-held conference. The conference schedule information of each to-be-held conference includes but is not limited to the conference theme, the location of the conference venue, the conference reporting expert, the conference time, the conference date, and the like. The to-be-held conference is a conference that the target user can attend and has not yet been held. The to-be-held conference can be an industry summit, an academic conference, and in other embodiments, the to-be-held conference can also be a corporate conference, and in addition, the to-be-held conference can also be a sub-conference under the same topic conference. The conference schedule knowledge graph includes different schedule information entities related to the to-be-held conference and the relationship between different schedule information entities. The schedule information entity is a single schedule information related to the to-be-held conference. Exemplarily, the conference schedule knowledge graph includes: schedule information entity 1 (expert Wang so-and-so), schedule information entity 2 (about xx industry summit), if the schedule information entity 1 and the schedule information entity 2 exist in association, then the conference schedule knowledge graph also includes the association between the schedule information entity 1 and the schedule information entity 2, that is, the expert Wang so-and-so is the reporter of the “about xx industry summit”. In addition, the first query text information is the query information input by the target user to understand the schedule information of the to-be-held conference.

[0064] Further, the execution subject of the information interaction method based on the conference schedule disclosed in the embodiments of the present application is a server, the server is wirelessly connected with the terminal of the target user, the terminal is a personal computer or a smart phone, the terminal is installed with an application program related to the conference information query, the server is a background server of the application program, and specifically can be an independent physical server or a cluster composed of multiple physical servers. An implementation scenario is that the target user opens the application program related to the conference information query through the terminal, and then inputs the query information in the input box in the main interface of the application program. The application program sends the query information to the server through a software development kit (Software Development Kit, SDK), and the server finally obtains the first query text information. In other embodiments, the target user can also input voice information related to the conference schedule through the terminal, and send the voice information to the server. The server converts the voice information into text information through a preset automatic speech recognition (Automatic Speech Recognition, ASR) model, and finally obtains the first query text information. Further, the server generates a reply result matched with the query demand of the target user based on the first query text, and returns the reply result to the terminal. For details, see Figure 2 .

[0065] Further, a feasible way to obtain the conference schedule knowledge graph is that the conference organizer inputs a set of related information of each to-be-held conference into the server through a corresponding terminal, the set of related information including schedule information of different to-be-held conferences and related theme label words. The server imports the schedule information into a preset Neo4j through Python, obtains the conference schedule knowledge graph and stores it. After obtaining the first query text information of the target user, the conference schedule knowledge graph is called from the Neo4j. It should be noted that Neo4j is a globally leading native graph database specially designed for storing, querying and processing data with complex association relationships.

[0066] S102: input the first query text information into a preset general large model, and determine the query requirement information of the target user for the conference schedule, the query requirement information including at least one target schedule information entity and a target query intention.

[0067] Specifically, the first query text information is input into the preset general large model to obtain the query requirement information, so as to realize the identification of the entity in the first query text and the query intention of the target user. The general large model can adopt DeepSeek or ChatGPT. The query requirement information reflects the query requirement of the target user for the schedule information of the to-be-held conference. In addition, the target schedule information entity is the schedule information entity identified based on the first query text information. The target query intention reflects the specific dimension information that the target user wants to know about the schedule information of the to-be-held conference. For example, the query requirement information includes “expert Li so-and-so” and “where is the location of the conference”, “expert Li so-and-so” is the schedule information entity, and “where is the location of the conference” is the specific query intention of the target user, i.e., the target query intention.

[0068] S103: if the target conference theme is not included in each target schedule information entity, then according to each target schedule information entity, query the associated schedule information entity from the conference schedule knowledge graph.

[0069] Specifically, the target conference theme is the specific conference theme of the to-be-held conference, reflecting the specific conference that the target user wants to know about the schedule information. If the target conference theme is not included in each target schedule information entity, it means that the first query text information does not involve the related content of the conference theme, and then according to the target schedule information entity, the associated schedule information entity is queried from the conference schedule knowledge graph, i.e., the schedule information related to each target schedule information entity belongs to the same to-be-held conference.

[0070] S104: According to the association of the schedule information entity and the target query intention, a first reply text information corresponding to the first query text information is generated, and the first reply text information is sent to the terminal of the target user.

[0071] Specifically, after the association of the schedule information entity is determined, the association of the schedule information entity and the target query intention is input into a preset large language model (LLM). The LLM filters the association of the schedule information entity that meets the target query intention from each association of the schedule information entity based on the target query intention, and uses the association of the schedule information entity as a reference for reply generation. For example, the target query intention is to query the venue location, and each association of the schedule information entity is the information of the reporting expert, the start time of the meeting, and the location information of the meeting venue. Then, the location information of the meeting venue is filtered. Finally, the LLM generates the first reply text information based on the association of the schedule information entity that meets the target query intention, and sends the first reply text information to the terminal of the target user, thereby providing the target user with reply information related to the schedule information that is more matched to the query demand of the target user. It should be noted that the LLM generates the first reply text information by using a knowledge alignment mechanism to constrain the authenticity of the reply to avoid LLM hallucinations. In addition, the LLM is a deepseek model or a Llama model that is injected with conference scene rules through instruction fine-tuning. This is prior art and will not be described here.

[0072] S105: If the target conference theme is included in each target schedule information entity, the target conference theme is reasonably verified. After the verification is passed, a second reply text information corresponding to the first query text information is generated according to the target conference theme and the target query intention, and the second reply text information is sent to the terminal of the target user.

[0073] Specifically, if the target conference theme is included in each target schedule information entity, in order to more accurately determine the conference that the target user wants to know about the schedule information, the target conference theme is reasonably checked. One implementable embodiment is: based on the above-mentioned schedule query record of the application program related to the conference information, the theme tag words of the multiple conferences that the historical user has queried the schedule information of are obtained. The theme tag word is a highly condensed keyword of the core theme, attribute, scene or value of the conference. Exemplarily, the conference theme is "Global Artificial Intelligence Industry Application Summit", and the theme tag words involved are "AI", "manufacturing industry", "smart city" and "industrial AI solution" and the like. Then, the occurrence frequency of each theme tag word in all theme tag words is counted. If the occurrence frequency exceeds a preset frequency threshold, it means that the occurrence frequency of the theme tag word is high, and the theme tag word is determined as the target tag word, that is, the theme tag word that the historical user is easy to involve in the conference. It should be noted that the schedule query record includes but is not limited to the theme tag words involved in the conference queried by the user of different user portraits and the keywords input by the user when querying and the like. In addition, the user portrait of the historical user is consistent with the user portrait of the target user. The user portrait is a structured and labeled user model based on the basic information of the user. The basic information includes but is not limited to the age, gender, occupation, hobby and the like of the user, which can be obtained through the registration information of the user on the application program under the premise of the consent or authorization of the user. One feasible way to determine the user portrait is to input the basic information of the user into a preset portrait construction model to obtain the corresponding user portrait. The portrait construction model can be a trained decision tree model or a support vector machine. The training process is briefly described as follows: the basic information sample data labeled with correct user portraits are divided into a training set, a test set and a validation set, and the model is trained with the three data sets. The training process is supervised by the cross-entropy loss function until the model converges. This is a prior art and will not be described here.

[0074] Further, based on the above-mentioned schedule query record, the multiple historical keywords input by the historical user when querying the schedule information of the target conference are obtained. The target conference is a conference involving a single target tag word. The occurrence number of each historical keyword in all historical keywords is counted. If the occurrence number exceeds a preset number threshold, the corresponding historical keyword is determined as the target keyword, that is, the historical keyword that the historical user is easy to input. For details, see Figure 3 .

[0075] Then, a first weight value of each target label word is determined, the first weight value being a ratio of the occurrence frequency of the target label word to a sum of occurrence frequencies of all target label words, the greater the first weight value, the greater the possibility that the conference concerned by the historical user or the target user is related to the target label word. A second weight value of each target keyword corresponding to the target label word is determined, the second weight value being a ratio of the occurrence times of the target keyword to a sum of occurrence times of all target keywords. Exemplarily, there are target label word a, target label word b and target label word c, the occurrence frequency of the target label word a is 20 times, the occurrence frequency of the target label word b is 30 times, and the occurrence frequency of the target label word c is 50 times, then the first weight value of the target label word a is: 20 times / (30 times+20 times+50 times)=0.2, the target label word a corresponds to target keyword A, target keyword B and target keyword C, the occurrence times of the target keyword A is 50 times, the occurrence times of the target keyword B is 30 times, and the occurrence times of the target keyword C is 20 times, then the second weight value of the target keyword A corresponding to the target label word a is: 50 times / (30 times+20 times+50 times)=0.5.

[0076] Further, at least one actual label word related to the target conference theme is screened from the subject label words in the above-mentioned related information set, and then at least one actual keyword is extracted from the first query text information through a preset TF-IDF algorithm, and the identified target conference theme is reasonably verified according to the actual label word, the actual keyword, the first weight value and the second weight value, and one implementable embodiment is:

[0077] The actual label word consistent with the target label word is determined as a key label word, and at least one actual keyword existing in each target keyword corresponding to the key label word is determined as a key keyword. The first product of the first weight value of at least one key label word and the second weight value of each key keyword is calculated, the greater the first product, the greater the possibility that the target user wants to query the to-be-held conference related to the key label word when inputting a single key keyword. The sum of each first product is obtained to obtain a first comprehensive result, the greater the first comprehensive result, the greater the overall possibility that the target user wants to query the to-be-held conference related to all key label words when inputting the first query text information. If the first comprehensive result is greater than a preset first threshold value, it indicates that the overall possibility that the target user wants to query the to-be-held conference related to all key label words is greater, and then it indicates that the possibility that the target user wants to query the conference of the target conference theme is greater, and then the reasonableness verification of the target conference theme is passed; otherwise, if the first comprehensive result is not greater than the first threshold value, it is determined that the reasonableness verification of the target conference theme is not passed.

[0078] After the verification passes, the schedule information related to the to-be-held meeting of the target meeting theme is screened from the meeting schedule knowledge graph according to the target meeting theme, and then the screened schedule information and the target query intention are input into a preset large language model to generate second reply text information corresponding to the first query text information, and the second reply text information is sent to the terminal of the target user. For details, see step S104, which is not repeated here.

[0079] In other embodiments, when the verification fails, it indicates that the target meeting theme identified based on the first query text information is biased and needs to be adjusted and optimized. A feasible adjustment and optimization method is as follows: if there is at least one actual keyword in each target keyword corresponding to the target label word, the target label word is determined as an important label word, and the at least one actual keyword is determined as an important keyword. Then, the first weight of the at least one important label word and the second weight of each important keyword are calculated. The greater the second product, the greater the possibility that the target user wants to query the to-be-held meeting involving the important label word when inputting a single important keyword. Then, the second products are summed to obtain a second comprehensive result corresponding to the important label word. The greater the second comprehensive result, the greater the possibility that the target user wants to query the to-be-held meeting involving the important label word when inputting the first query text information. If the second comprehensive result is greater than a preset second threshold, it indicates that the possibility that the target user wants to query the to-be-held meeting involving the corresponding important label word is greater, and then the corresponding important label word is determined as a to-be-selected label word.

[0080] Further, each to-be-selected label word is sent to the terminal of the target user for selection. If at least one selected label word (a label word selected by the target user from each to-be-selected label word) is received from the terminal, the number of selected label words involved in each to-be-held meeting is counted. The greater the number, the greater the possibility that the corresponding to-be-held meeting is the meeting that the target user wants to query. Then, the target meeting theme is adjusted and optimized according to the number and the selected label words involved to obtain a final meeting theme. The specific process is as follows: the second comprehensive results corresponding to the selected label words involved are summed to obtain a third comprehensive result. The greater the third comprehensive result, the greater the possibility that the target user wants to query the to-be-held meeting corresponding to the selected label words involved.

[0081] Further, a result correction coefficient corresponding to the number is determined from a preset coefficient matching table. The larger the number, the larger the result correction coefficient. The result correction coefficient is not less than 1. The coefficient matching table includes different number ranges and corresponding result correction coefficients. For example, the number range is 1-3, and the corresponding result correction coefficient is 1.2. The number range is 3-5, and the corresponding result correction coefficient is 1.4. If the number is 4, the to-be-held meeting involves 4 selected label words, and is within the number range 3-5. Then, the corresponding result correction coefficient is 1.4.

[0082] The result correction coefficient is multiplied by the third comprehensive result to obtain a corrected result corresponding to the single to-be-held meeting. If the corrected result is greater than a preset third threshold value, it indicates that the target user wants to query the corresponding to-be-held meeting with a higher probability. Then, the meeting theme of the to-be-held meeting corresponding to the corrected result is determined as the final meeting theme. Finally, according to the final meeting theme and the target query intent, the third reply text information corresponding to the first query text information is generated, that is, the final meeting theme and the target query intent are input into the large language model to generate the third reply text information. For details, see step S104, which is not repeated here. At the same time, the third reply text information is sent to the terminal of the target user.

[0083] In yet another embodiment, when the verification passes, it indicates that the target meeting theme is correct, and the to-be-held meeting that the target user wants to query is correct. Then, it is detected whether the target user is in an input state. If yes, it indicates that the target user wants to query other related information about the to-be-held meeting of the target meeting theme. Then, the actual label word consistent with the target label word is determined as a key label word. A third product of the first weight value of at least one key label word and the second weight value of each target keyword corresponding to the key label word is calculated. The larger the third product, the greater the possibility that the target user wants to understand the schedule information of the to-be-held meeting of the target meeting theme and input the corresponding target keyword. Then, each third product corresponding to the same target keyword is summed to obtain a target comprehensive result. The larger the target comprehensive result, the greater the possibility that the target user inputs the corresponding target keyword. If the target comprehensive result is greater than a preset fourth threshold value, the corresponding target keyword is determined as a candidate keyword.

[0084] The sub-keyword currently input by the target user is acquired in real time. The candidate keyword containing the sub-keyword is displayed in the input box in the terminal in the form of prompt text. The target user can directly click the displayed prompt text instead of continuing to manually input, which better improves the experience of the target user in inputting query information.

[0085] In another embodiment, if the first query text information of the target user for the conference schedule is not acquired after a preset time period, it indicates that the target user is difficult to select the desired to-be-held conference from the plurality of to-be-held conferences, and therefore, under the premise that the target user agrees or authorizes, a collection pop-up window of the location information and the meeting attendance departure time is sent to the terminal to finally acquire the current location and the meeting attendance departure time of the target user, wherein the meeting attendance departure time is the expected departure time of the target user to attend the to-be-held conference. Then, according to the schedule information of each to-be-held conference input by the conference organizer, the venue locations of the to-be-held conferences are acquired.

[0086] The current location is set as a starting point, and each venue location is set as a terminal point. The expected time length of the target user to reach each to-be-held conference venue is determined through a preset navigation tool, and the meeting attendance departure time is added to the expected time length to obtain the expected arrival time. Then, if the start time of the to-be-held conference corresponding to the to-be-held conference is later than the corresponding expected arrival time, it indicates that the target user can attend the to-be-held conference on time, and the to-be-held conference is determined as a to-be-recommended conference. Finally, the first weights of the target label words involved in the to-be-recommended conference are summed to obtain a weight sum, and if the weight sum is greater than a preset weight threshold, it indicates that the target user wants to attend the to-be-recommended conference with a higher possibility, and the schedule information of the to-be-recommended conference is directly sent to the terminal of the target user.

[0087] The implementation principle of the information interaction method of the conference schedule based on the embodiments of the present application is as follows: based on the first query text information, a general large model is used to analyze and determine the query demand of the target user for the to-be-held conference, and then when each target schedule information entity does not contain the target conference theme, the associated schedule information entity associated with the target schedule information entity is quickly queried through the relationship between the entities in the conference schedule knowledge graph, and is used as the reference content for generating the reply corresponding to the first query text information, and then the first reply text information that meets the demand of the target user is generated; when each target schedule information entity contains the target conference theme, in order to more accurately determine the to-be-held conference that the target user wants to query the schedule information, the target conference theme is reasonably verified, and after the verification is passed, the second reply text information related to the to-be-held conference of the target conference theme is generated in combination with the query intention of the target user (target query intention), so that the matching of the queried conference schedule information and the demand of the target user is improved.

[0088] The following is an apparatus embodiment of the present application, which can be used to execute the method embodiments of the present application. For details not disclosed in the apparatus embodiments of the present application, please refer to the method embodiments of the present application.

[0089] Please refer to Figure 4A structural schematic diagram of the information interaction device based on the conference schedule provided by the embodiments of the present application. The information interaction device based on the conference schedule can be realized by software, hardware or a combination of both to become all or part of the device. The device includes an information acquisition module 11, a demand determination module 12, an entity query module 13, a first reply module 14 and a second reply module 15.

[0090] The information acquisition module 11 is configured to acquire first query text information of a target user for a conference schedule and acquire a conference schedule knowledge graph. The conference schedule knowledge graph includes different schedule information entities related to a to-be-held conference and relationships between different schedule information entities. The schedule information entity is a single schedule information related to the to-be-held conference.

[0091] The demand determination module 12 is configured to input the first query text information into a preset general large model to determine query demand information of the target user for the conference schedule. The query demand information includes at least one target schedule information entity and a target query intent.

[0092] The entity query module 13 is configured to, if the target conference theme is not included in each target schedule information entity, query associated schedule information entities from the conference schedule knowledge graph according to each target schedule information entity. The associated schedule information entities belong to the same schedule information related to the to-be-held conference as each target schedule information entity.

[0093] The first reply module 14 is configured to generate first reply text information corresponding to the first query text information according to the associated schedule information entity and the target query intent, and send the first reply text information to a terminal of the target user.

[0094] The second reply module 15 is configured to, if the target conference theme is included in each target schedule information entity, perform rationality verification on the target conference theme. After the verification passes, generate second reply text information corresponding to the first query text information according to the target conference theme and the target query intent, and send the second reply text information to the terminal of the target user.

[0095] Optionally, the second reply module 15 is specifically configured to:

[0096] acquire theme label words of a plurality of conferences for which the historical user has queried schedule information, and determine at least one target label word according to each theme label word. The target label word is a theme label word that the historical user pays attention to and is easy to involve in the conference. The historical user has the same user portrait as the target user.

[0097] The input historical keywords are obtained when a historical user queries schedule information of a target meeting, and at least one target keyword is determined according to each historical keyword, the target keyword is a historical keyword that is easy for the historical user to input, and the target meeting is a meeting related to a single target label keyword;

[0098] A first weight value of each target label keyword is determined, and a second weight value of a target keyword corresponding to each target label keyword is determined, the first weight value represents a possibility of a meeting that is focused on by the historical user being related to the target label keyword, and the second weight value represents a possibility of the target keyword being input by the historical user when querying schedule information of the target meeting;

[0099] At least one actual label keyword related to the target meeting theme is obtained, at least one actual keyword is extracted from the first query text information, and the target meeting theme is reasonably verified according to the actual label keyword, the actual keyword, the first weight value and the second weight value.

[0100] Optionally, the second reply module 15 is specifically used for:

[0101] An actual label keyword consistent with the target label keyword is determined as a key label keyword, and at least one actual keyword existing in each target keyword corresponding to the key label keyword is determined as a key keyword;

[0102] A first product of the first weight value of at least one key label keyword and the second weight value of each key keyword is calculated, and each first product is summed to obtain a first comprehensive result;

[0103] The first comprehensive result is compared with a preset first threshold value, if the first comprehensive result is greater than the first threshold value, it is determined that the reasonable verification of the target meeting theme is passed;

[0104] If the first comprehensive result is not greater than the first threshold value, it is determined that the reasonable verification of the target meeting theme is not passed.

[0105] Optionally, as shown in Figure 5 The device further includes a third reply module 16, which is specifically used for:

[0106] When the verification is not passed, if an actual keyword exists in each target keyword corresponding to the target label keyword, the corresponding target label keyword is determined as an important label keyword, and at least one actual keyword existing in each target keyword corresponding to the important label keyword is determined as an important keyword;

[0107] A second product of the first weight value of at least one important label keyword and the second weight value of each important keyword is calculated, each second product is summed to obtain a second comprehensive result of the corresponding important label keyword;

[0108] If the second comprehensive result is greater than a preset second threshold value, the corresponding important label word is determined as a to-be-selected label word, each to-be-selected label word is sent to the terminal, and at least one selected label word sent by the terminal is received;

[0109] The number of selected label words involved in each to-be-held meeting is counted, and the target meeting theme is adjusted and optimized according to the number and the selected label words involved, to obtain a final meeting theme;

[0110] According to the final meeting theme and the target query intention, third reply text information corresponding to the first query text information is generated, and the third reply text information is sent to the terminal of the target user.

[0111] Optionally, the third reply module 16 is specifically configured to:

[0112] The second comprehensive results corresponding to the selected label words involved are summed to obtain a third comprehensive result;

[0113] According to the number, a result correction coefficient is determined, the greater the number is, the greater the result correction coefficient is, and the result correction coefficient is a positive number not less than 1;

[0114] The result correction coefficient is multiplied by the third comprehensive result to obtain a corrected result, and if the corrected result is greater than a preset third threshold value, the meeting theme of the to-be-held meeting corresponding to the corrected result is determined as the final meeting theme.

[0115] Optionally, the device further includes an input auxiliary module 17, which is specifically configured to:

[0116] After the verification passes, if it is detected that the target user is in an inputting state, the actual label word consistent with the target label word is determined as a key label word, and a third product of a first weight value of at least one key label word and a second weight value of each target keyword corresponding to the key label word is calculated;

[0117] The third products corresponding to the same target keyword are summed to obtain a target comprehensive result, and if the target comprehensive result is greater than a preset fourth threshold value, the corresponding target keyword is determined as a to-be-selected keyword;

[0118] The sub-keyword currently input by the target user is obtained, and the to-be-selected keyword containing the sub-keyword is displayed in the input box in the terminal in the form of a prompt text.

[0119] Optionally, the device further includes a meeting recommendation module 18, which is specifically configured to:

[0120] If the first query text information of the target user for the meeting schedule is not obtained after a preset time period, the current position of the target user, the meeting departure time, and the meeting location of each to-be-held meeting are obtained;

[0121] According to the current location and the location of each meeting site, a predicted time length of the target user to reach the meeting site is determined, and according to the meeting attendance departure time and the predicted time length, a predicted arrival time of the target user to reach the meeting site where each to-be-held meeting is located is determined;

[0122] If the meeting start time corresponding to the to-be-held meeting is later than the corresponding predicted arrival time, the to-be-held meeting corresponding to the meeting start time is determined as the to-be-recommended meeting;

[0123] The first weight values of the target label words involved in the to-be-recommended meeting are summed to obtain a weight sum, and if the weight sum is greater than a preset weight threshold, the schedule information of the to-be-recommended meeting is sent to the terminal.

[0124] It should be noted that the above embodiment provides a conference schedule-based information interaction device, which, when performing the conference schedule-based information interaction method, is only exemplified by the division of the above functional modules. In actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the above-described functions. In addition, the conference schedule-based information interaction device and the conference schedule-based information interaction method embodiment provided by the above embodiment belong to the same concept, and the implementation process is detailed in the method embodiment, which will not be described here.

[0125] The application also discloses a computer readable storage medium, and the computer readable storage medium stores a computer program, wherein the computer program is executed by a processor to realize the conference schedule-based information interaction method of the above embodiment.

[0126] The computer program can be stored in the computer readable medium, the computer program includes computer program code, the computer program code can be in the form of source code, object code, executable file or some middleware form, etc., and the computer readable medium includes any entity or device, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. that can carry computer program code. It should be noted that the computer readable medium includes but is not limited to the above components.

[0127] The computer readable storage medium stores the conference schedule-based information interaction method of the above embodiment in the computer readable storage medium, and is loaded and executed on the processor to facilitate the storage and application of the above method.

[0128] The embodiment of the application further discloses an electronic device, a computer program is stored in a computer readable storage medium, and the computer program is loaded and executed by a processor to realize the information interaction method based on a conference schedule.

[0129] The electronic device can be a desktop computer, a notebook computer, or a cloud server, and the electronic device includes but is not limited to a processor and a memory, for example, the electronic device can further include an input / output device, a network access device, and a bus.

[0130] The processor can be a central processing unit (CPU), and can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. The general-purpose processor can be a microprocessor or any conventional processor, and the application does not limit the processor.

[0131] The memory can be an internal storage unit of the electronic device, for example, a hard disk or a memory of the electronic device, or an external storage device of the electronic device, for example, a plug-in hard disk, a smart memory card (SMC), a secure digital card (SD), or a flash memory card (FC) equipped on the electronic device, and the memory can also be a combination of the internal storage unit and the external storage device of the electronic device. The memory is used to store computer programs and other programs and data required by the electronic device, and can also be used to temporarily store data that has been output or will be output, and the application does not limit the memory.

[0132] The electronic device stores the information interaction method based on the conference schedule in the memory of the electronic device, and loads and executes the information interaction method based on the conference schedule on the processor of the electronic device, which is convenient to use.

[0133] The above is only an exemplary embodiment of the disclosure, and cannot limit the scope of the disclosure. Any equivalent changes and modifications made in accordance with the teachings of the disclosure are still within the scope of the disclosure. The application intends to cover any variations, uses, or adaptive changes of the disclosure, which follow the general principles of the disclosure and include common knowledge or conventional technical means in the technical field not recorded in the disclosure. The scope and spirit of the disclosure are defined by the claims.

Claims

1. A method for information interaction based on meeting schedule, characterized in that, The method comprises: acquiring first query text information of a target user for a conference schedule, and acquiring a conference schedule knowledge graph, the conference schedule knowledge graph comprising different schedule information entities related to a to-be-held conference and relationships between different schedule information entities, the schedule information entity being a single schedule information related to the to-be-held conference; inputting the first query text information into a preset general large model to determine query demand information of the target user for the conference schedule, the query demand information comprising at least one target schedule information entity and a target query intent; if none of the target schedule information entities contains a target conference theme, querying, according to each target schedule information entity, an associated schedule information entity from the conference schedule knowledge graph, the associated schedule information entity belonging to the same schedule information related to the to-be-held conference as each target schedule information entity; generating first reply text information corresponding to the first query text information according to the associated schedule information entity and the target query intent, and sending the first reply text information to a terminal of the target user. If the target conference theme is included in each of the target schedule information entities, the target conference theme is subjected to rationality verification, including: obtaining theme label words of a plurality of conferences whose schedule information is queried by a historical user, and determining at least one target label word according to each of the theme label words, the target label word being a theme label word that is easily involved in a conference that is focused on by the historical user, the historical user being consistent with a user portrait of the target user; obtaining historical keywords input by the historical user when querying schedule information of a target conference, and determining at least one target keyword according to each of the historical keywords, the target keyword being a historical keyword that is easily input by the historical user, the target conference being a conference involving a single target label word; determining a first weight of each target label word, and determining a second weight of a target keyword corresponding to each target label word, the first weight representing a possibility size of a conference that is focused on by the historical user involving the target label word, and the second weight representing a possibility size of the historical user inputting the target keyword when querying schedule information of the target conference; obtaining at least one actual label word involved in a conference corresponding to the target conference theme, and extracting at least one actual keyword from the first query text information, and performing rationality verification on the target conference theme according to the actual label word, the actual keyword, the first weight, and the second weight; after the verification passes, generating second reply text information corresponding to the first query text information according to the target conference theme and the target query intent, and sending the second reply text information to a terminal of the target user; if the verification fails, if the actual keyword exists in each target keyword corresponding to the target label word, determining the corresponding target label word as an important label word, and determining at least one actual keyword existing in each target keyword corresponding to the important label word as an important keyword; calculating a second product of the first weight of at least one important label word and the second weight of each important keyword, summing each second product to obtain a second comprehensive result of the corresponding important label word; if the second comprehensive result is greater than a preset second threshold, determining the corresponding important label word as a candidate label word, sending each candidate label word to the terminal, and receiving at least one selected label word sent by the terminal; counting a number of selected label words involved in each conference to be held, and adjusting and optimizing the target conference theme according to the number and the selected label words involved, to obtain a final conference theme; generating third reply text information corresponding to the first query text information according to the final conference theme and the target query intent, and sending the third reply text information to the terminal of the target user. After the verification passes, if it is detected that the target user is in an inputting state, an actual label word consistent with the target label word is determined as a key label word, a third product of a first weight value of at least one key label word and a second weight value of each target keyword corresponding to the key label word is calculated, each third product corresponding to the same target keyword is summed to obtain a target comprehensive result, if the target comprehensive result is greater than a fourth threshold value, the corresponding target keyword is determined as a candidate keyword, a sub-keyword currently input by the target user is obtained, and the candidate keyword containing the sub-keyword is displayed in the form of prompt text in an input box in the terminal.

2. The information interaction method based on the conference schedule according to claim 1, wherein, The rationality verification on the target conference theme according to the actual label word, the actual keyword, the first weight value and the second weight value specifically includes: determining an actual label word consistent with the target label word as a key label word, and determining at least one actual keyword existing in each target keyword corresponding to the key label word as a key keyword; calculating a first product of the first weight value of at least one key label word and the second weight value of each key keyword, and summing each first product to obtain a first comprehensive result; comparing the first comprehensive result with a first threshold value, if the first comprehensive result is greater than the first threshold value, it is determined that the rationality verification on the target conference theme passes; if the first comprehensive result is not greater than the first threshold value, it is determined that the rationality verification on the target conference theme fails.

3. The information interaction method based on conference schedule according to claim 1, characterized in that, The adjustment and optimization on the target conference theme according to the number and the selected label word involved specifically includes: summing the second comprehensive results corresponding to the selected label words involved to obtain a third comprehensive result; determining a result correction coefficient according to the number, the larger the number is, the larger the result correction coefficient is, and the result correction coefficient is a positive number not less than 1; multiplying the result correction coefficient and the third comprehensive result to obtain a corrected result, if the corrected result is greater than a third threshold value, the conference theme of the meeting to be held corresponding to the corrected result is determined as the final conference theme.

4. The information interaction method based on conference schedule according to claim 1, characterized in that, The method further includes: if the first query text information of the target user for the conference schedule is not obtained after a preset time period, the current location of the target user, the departure time for attending the meeting and the location of each meeting venue of the meetings to be held are obtained; determining the expected time length for the target user to arrive at the meeting venue according to the current location and the location of each meeting venue, and determining the expected arrival time of the target user to arrive at each meeting venue of the meetings to be held according to the departure time for attending the meeting and the expected time length; if the start time of the corresponding meeting to be held is later than the corresponding expected arrival time, the corresponding meeting to be held is determined as a meeting to be recommended. Sum the first weight values of the target label words related to the conference to be recommended to obtain a weight sum, and if the weight sum is greater than a preset weight threshold, schedule information of the conference to be recommended is sent to the terminal.

5. An information interaction apparatus based on a conference schedule, for implementing the information interaction method based on a conference schedule according to any one of claims 1 to 4, characterized in that, Comprise: An information acquisition module (11) is configured to acquire first query text information of a target user for a conference schedule and acquire a conference schedule knowledge graph, the conference schedule knowledge graph comprising different conference information entities related to different conferences to be held and relationships between different conference information entities, and the conference information entity being a single conference information related to the conference to be held; A demand determination module (12) is configured to input the first query text information into a preset general large model to determine query demand information of the target user for the conference schedule, the query demand information comprising at least one target conference information entity and a target query intent; An entity query module (13) is configured to, if the target conference theme is not contained in each target conference information entity, query associated conference information entities from the conference schedule knowledge graph according to each target conference information entity, the associated conference information entities belonging to the same conference information related to the conference to be held as each target conference information entity; A first reply module (14) is configured to generate first reply text information corresponding to the first query text information according to the associated conference information entities and the target query intent, and send the first reply text information to a terminal of the target user; A second reply module (15) is configured to, if the target conference theme is contained in each target conference information entity, perform rationality verification on the target conference theme, and after the verification passes, generate second reply text information corresponding to the first query text information according to the target conference theme and the target query intent, and send the second reply text information to the terminal of the target user.

6. A computer-readable storage medium having stored therein a computer program, characterized in that, The computer program is loaded and executed by the processor, and the method of any one of claims 1-4 is realized.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, The processor loads and executes the computer program, and the method of any one of claims 1-4 is realized.

Images