Method of determining interaction information, device, electronic apparatus, storage medium and computer program

Abstract

To provide a method of determining interaction information, a device, an electronic apparatus, a storage medium and a computer program which can create questioning dimensions for query information by combining large model previous specialized knowledge and subsequent supplements in user historical query information by means of the method of the chain of thoughts.SOLUTION: A method is to determine a plurality of questioning dimensions according to query information of a subject and historical query information, where each questioning dimension includes a dimension name and a plurality of options. The method is also to determine a target questioning dimension from the plurality of questioning dimensions according to evaluation values of the plurality of questioning dimensions and whether semantic information of the plurality of questioning dimensions are consistent with semantic information of a query result associated with the query information; and determine the interaction information according to the dimension name and the plurality of options in the target questioning dimension.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present disclosure relates to the field of artificial intelligence technology, in particular to the fields of large-scale models, generative models, NLP, intelligent search, etc., and more specifically, the present disclosure provides a method, device, electronic device, storage medium, and computer program for identifying interaction information. [Background technology]

[0002] In complex search scenarios, due to the diversity and ambiguity of user expressions, search engines cannot accurately capture the actual search intent, and there exists a gap between the search results and the actual needs of users. Summary of the Invention

[0003] The present disclosure provides a method, a device, an electronic device, a storage medium, and a computer program for identifying interaction information.

[0004] According to one aspect of the present disclosure, a method for identifying interaction information is provided, the method including: identifying a plurality of question dimensions based on target query information and historical query information, each question dimension including a dimension name and a plurality of options; identifying a target question dimension from the plurality of question dimensions based on whether semantic information of the plurality of question dimensions matches with semantic information of a query result related to the query information and evaluation values ​​of the plurality of question dimensions; and identifying interaction information based on the dimension name and the plurality of options in the target question dimension.

[0005] According to another aspect of the present disclosure, a device for identifying interaction information is provided, the device including a first identifying module, a second identifying module, and a third identifying module. The first identifying module identifies a plurality of question dimensions based on target query information and historical query information, each of the question dimensions including a dimension name and a plurality of options. The second identifying module identifies a target question dimension from the plurality of question dimensions based on whether semantic information of the plurality of question dimensions matches with semantic information of a query result related to the query information, and on evaluation values ​​of the plurality of question dimensions. The third identifying module identifies interaction information based on the dimension name and the plurality of options in the target question dimension.

[0006] According to another aspect of the present disclosure, there is provided an electronic device including at least one processor and a memory communicatively coupled to the at least one processor, the memory storing instructions executable by the at least one processor, the instructions being executed by the at least one processor such that the at least one processor can perform a method provided in the present disclosure.

[0007] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium having computer instructions stored thereon is provided, where the computer instructions cause a computer to perform the methods provided in the present disclosure.

[0008] According to another aspect of the present disclosure, there is provided a computer program which, when executed by a processor, implements the methods provided in the present disclosure.

[0009] It should be understood that the contents described in this section are not intended to depict key points or important features of the embodiments of the present disclosure, and are not intended to limit the scope of the present disclosure. Other features of the present disclosure will be readily understood from the following description. [Brief description of the drawings]

[0010] The drawings are used for a better understanding of the present technical solution and are not intended to limit the present disclosure.

[0011] [Figure 1] FIG. 1 is a schematic diagram of an application scene of the method and apparatus for identifying interaction information according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a schematic flowchart of a method for identifying interaction information according to an embodiment of the present disclosure. [Figure 3A] FIG. 3A is a schematic principle diagram of a method for identifying interaction information according to an embodiment of the present disclosure. [Figure 3B] FIG. 3B is a schematic diagram of an interaction page according to an embodiment of the present disclosure. [Figure 4] FIG. 4 is a schematic block diagram of an apparatus for identifying interaction information according to an embodiment of the present disclosure. [Diagram 5] FIG. 5 is a structural block diagram of an electronic device implementing the method for identifying interaction information according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, exemplary embodiments of the present disclosure will be described with reference to the drawings. Hereinafter, various details of the embodiments of the present disclosure are included for easier understanding, and they should be considered as examples. Therefore, it should be understood that a person skilled in the art can make various changes and modifications to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and conciseness, the following description omits descriptions of well-known functions and configurations.

[0013] In the technical solution disclosed herein, the collection, storage, use, processing, transmission, provision, disclosure and other processing of user personal information are all in accordance with the provisions of relevant laws and regulations and are not contrary to public order and morals.

[0014] In the technical solutions disclosed herein, before obtaining or collecting personal information of a user, the permission or consent of the user is obtained in each case.

[0015] In complex search scenarios, due to the diversity and ambiguity of user expressions, search engines cannot accurately capture the actual search intent, and there exists a gap between the search results and the actual needs of users.

[0016] For example, the query information of "adjusting the body with Chinese herbal medicine" is a relatively complex and vague need, and the query information is related to the user's personalized preferences, but the search engine returns a wide range of query results related to keywords such as Chinese herbal medicine and body adjustment, which contain many contents that do not meet the user's needs and cannot fully meet the user's personalized needs. Some search engines use a related search method to recommend other possible query information to the user, but this method still has difficulty in discovering the user's true needs and is prone to inclusion of contents unrelated to the user's needs.

[0017] The embodiment of the present disclosure provides a method for identifying interaction information, which can combine large-scale model prior expertise and post-hoc supplementation in user history query information through a thought chain method to generate a query dimension for query information. By asking a user a question, the user is induced to provide more detailed information about the query target, and the query needs of the user are mined. In this way, the search engine can provide query results that match the needs of the user, improve the accuracy and relevance of the query results, and provide the user with a more personalized and satisfying search experience. In addition, the user can actively participate in the query process without passively receiving the content recommended by the search engine, thereby improving the user's activity and participation, and improving the user's usage experience.

[0018] Hereinafter, the technical solution of the present disclosure will be described in detail with reference to the drawings and specific embodiments.

[0019] FIG. 1 is a schematic diagram of an application scene of the method and apparatus for identifying interaction information according to an embodiment of the present disclosure.

[0020] Note that FIG. 1 is an example of a system architecture to which the embodiments of the present disclosure can be applied so that a person skilled in the art can understand the technical content of the present disclosure, and does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments, or scenes.

[0021] 1, the system architecture 100 according to the embodiment may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is a medium that provides a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as, for example, wired and / or wireless communication links.

[0022] Using terminal devices 101, 102, 103, users can interact with server 105 via network 104 to send and receive messages, etc. Terminal devices 101, 102, 103 can be various electronic devices that have a display and support web page browsing, including, but not limited to, smartphones, tablet computers, laptop portable computers, desktop computers, etc.

[0023] The server 105 may be a server that provides various services, for example, a background management server (just an example) that provides support for websites that users browse using the terminal devices 101, 102, 103. The background management server can perform processing such as analysis on data such as received user requests, and feed back the processing results (e.g., interaction information generated based on query information input by the user) to the terminal devices.

[0024] It should be noted that the method for identifying interaction information according to the embodiment of the present disclosure may generally be performed by the server 105. Accordingly, the device for identifying interaction information according to the embodiment of the present disclosure may generally be installed in the server 105. The method for identifying interaction information according to the embodiment of the present disclosure may be performed by a server or a server cluster different from the server 105 and capable of communicating with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the device for identifying interaction information according to the embodiment of the present disclosure may be installed in a server or a server cluster different from the server 105 and capable of communicating with the terminal devices 101, 102, 103 and / or the server 105.

[0025] It should be understood that the number of terminal devices, networks, and servers in Figure 1 is merely an example, and any number of terminal devices, networks, and servers may be included as needed.

[0026] FIG. 2 is a schematic flowchart of a method for identifying interaction information according to an embodiment of the present disclosure.

[0027] As shown in FIG. 2, the method 200 for identifying interaction information may include operations S210-S230.

[0028] In operation S210, a plurality of query dimensions are identified based on the target query information and the historical query information, where each query dimension includes a dimension name and a plurality of options.

[0029] For example, the audience includes users who use search engines.

[0030] For example, the query information may be a content that is input by a user and requires a query, and the data type of the query information may be text, for example, the query information is "interesting novel recommendations".

[0031] For example, the historical query information may include query information within a predetermined period of time in the past, and the predetermined period may be half a month. In some embodiments, the content supplemented and expressed by the user may be the historical query information. For example, after the user inputs one query information into the search engine, the user may continue to input some other content without exiting the search engine, or may select the query content provided by some pages in the search engine, and these contents may be the historical query information.

[0032] For example, dimension names may include "novel type", "novel width", etc., options related to "novel type" may include "suspense mystery", "science fiction", etc., and options related to "novel width" may include "full length", "medium length", "short length", etc.

[0033] For example, keyword extraction may be performed on the query information and the historical query information, and a mapping relationship between the keywords and the question dimensions may be preset, and then the question dimensions may be identified based on the mapping relationship.

[0034] In operation S220, a target question dimension is identified from the multiple question dimensions based on whether the semantic information of the multiple question dimensions matches with the semantic information of the query result related to the query information and based on the evaluation values ​​of the multiple question dimensions.

[0035] For example, the query result may represent the content obtained by searching the query information. The data type of the query result may include text, image, video, etc. For example, the query result includes the content "the recommended full-length novel is the first novel, the recommended medium-length novel is the second novel, and the recommended short story is the third novel", and this query content is consistent with the semantics of "novel page length" in the question dimension.

[0036] For example, the evaluation value may represent the degree of influence of the question dimension on the query content, or may represent the degree of influence of the question dimension on the understanding of the user query needs. The evaluation value may be a specific numerical value, or may be an evaluation parameter representing the high or low evaluation value. The evaluation value corresponding to the question dimension may be preset, or may be evaluated using a large-scale model.

[0037] For example, a question dimension that does not match the query result semantics and has a high evaluation value may be identified as a target question dimension.

[0038] In operation S230, interaction information is identified based on the dimension names and multiple options in the target question dimension.

[0039] For example, the dimension name in the target question dimension and a plurality of options may be used as the interaction information. Also, for example, the dimension name in the target question dimension can be expressed in natural language, and the dimension name "novel type" can be converted to, for example, "What type of novel do you want to read?", and then the converted dimension information and options can be used as the interaction information, which makes it easier for the user to understand the interaction information.

[0040] The embodiment of the present disclosure mines the user's query needs by asking the user questions to induce the user to provide more detailed information about the query target. In this way, the search engine can provide query results that match the user's needs, improve the accuracy and relevance of the query results, and provide the user with a more personalized and satisfying search experience. In addition, the user can actively participate in the query process instead of passively receiving the content recommended by the search engine, thereby improving the user's activity and participation and improving the user's usage experience.

[0041] In practical applications, after the interaction information is obtained, the interaction information can be applied in multiple ways. In one application scenario, the interaction information is displayed through a front-end page, so that the user can make a selection operation based on the question dimension to supplement the query content. In another application scenario, the interaction information can be used to build a training sample, and the training sample can be used to train another model, so that the other model trained in the online application can be used to interact with the user. The resources required for the other model may be smaller than those required for the large-scale model, and the other model may be a Sparse Fine Tuning (SFT) model.

[0042] According to another embodiment of the present disclosure, the historical query information may be first filtered, and the filtering process may include identifying, among the candidate historical query information, candidate historical query information that satisfies a predetermined filtering condition as the historical query information, where the predetermined filtering condition may include at least one of: an intent of the candidate historical query information matches an intent of the query information; a search frequency of the candidate historical query information satisfies a predetermined frequency condition; and an entity included in the candidate historical query information matches an entity included in the query information.

[0043] For example, first, historical query information within a predetermined period in the past may be acquired, and this historical query information may be set as candidate historical query information.

[0044] Then, it can be determined whether the intention of the candidate history query information and the query information is consistent. For example, it can be determined whether the intention is consistent by a binary classification method, for example, the candidate history query information and the query information are input into a pre-trained classification model, and the classification model determines whether the intention is consistent. For example, it can also be determined whether the intention of the candidate history query information and the query information is consistent, by performing intention recognition on each of them.

[0045] It is understood that the frequency of the candidate history query information can then be calculated, and the candidate history query information can then be selected based on a predetermined frequency condition, and the frequency can reflect a user's preference. The predetermined frequency condition may include that the frequency of the candidate history query information is greater than a threshold, and the threshold may be a numerical value such as 5. The predetermined frequency condition may include that the frequency ranking of the candidate history query information is in a top predetermined ranking, for example, the frequency is in the top 10.

[0046] Next, determine whether the entities included in the selected candidate historical query information match the entities included in the query information, and the manner of identifying the entities may include named entity identification, and this embodiment is not limited thereto.

[0047] For example, the query information and the candidate history query information are, respectively, “how to repair a first brand of mobile phone if it breaks down” and “how to repair a second brand of mobile phone if it breaks down”, and although the intents of the two are consistent, the entities are not consistent.

[0048] This embodiment can select historical query information that is highly relevant to the query information and more interesting to the user from the candidate historical query information based on predetermined selection conditions, thereby ensuring the accuracy of the question dimension identified based on the subsequent historical query information.

[0049] FIG. 3A is a schematic principle diagram of a method for identifying interaction information according to an embodiment of the present disclosure, and FIG. 3B is a schematic diagram of an interaction page according to an embodiment of the present disclosure.

[0050] According to another embodiment of the present disclosure, a method for identifying interaction information includes operations S310-S340, where operation S340 is performed before operation S310.

[0051] In operation S340, it is determined whether a question is required.

[0052] In operation S310, a number of query dimensions are identified.

[0053] In operation S320, a target query dimension is selected.

[0054] In operation S330, interaction information is identified.

[0055] In the above operation S340, the determination of whether a question is required is a pre-processing process, which mainly determines whether there is a complex need and whether the query information needs to be supplemented by the way of the question. The pre-processing presentation information template can be used to determine whether a question is required. For example, the query information input by the user and the pre-processing presentation information template can be combined and input into a large-scale model, and the determination result can be output by the large-scale model. The pre-processing presentation information template can be pre-arranged and stored in a predetermined storage area, and can be read out from the storage area as needed.

[0056] For example, the pre-processing presentation information template may include a thought chain, which includes multiple steps, and the thought chain instructs the large-scale model to reason according to the steps in the thought chain. For example, the multiple steps included in the thought chain are as follows: The essence of the analysis is to determine whether the query result needs to consider different context classification or user personalized analysis, such as personalized preferences, user personal images, etc. If so, determine that the current query belongs to complex needs and needs to be asked; otherwise, determine that there is no need to ask.

[0057] In addition, for example, the pre-processing presentation information template may include examples, and examples are mainly used to explain which is a personalized situation and which is a classification of different contexts. For example, as an example in the pre-processing presentation information template, the query information is "interesting novel recommendations", and when answering this question, it is necessary to consider personalized factors such as the user's reading habits and favorite literary types. Take the query information "Beijing house prices" as an example, since there is a large difference in the prices of houses in different regions and different types, the answer involves specific context factors such as regions and house types. These questions belong to complex needs and are open to questions. Conversely, take the query information "what time is 1 o'clock" as an example, such a need is clear and does not change according to the context or the specific situation of the individual, so it does not belong to complex needs and does not need to be asked.

[0058] This embodiment ensures the accuracy of the questions by determining the question needs, so that, when appropriate, the questions can obtain more user information and provide more accurate personalized answers.

[0059] In the above operation S310, in the process of identifying multiple question dimensions, multiple question dimensions can be identified based on the target query information and the historical query information. In addition, in the identification process, two angles, pre-refined dimension analysis and post-refined dimension mining, can be considered. For example, a presentation information template can be presented to a large-scale model, and the contextual elements and personalization elements that affect answering the query information can be decomposed from two angles, pre-refined and post-refined, to summarize a set of question dimensions. Options that can be refined according to the need dimensions are generated to form an interactive user selection interface.

[0060] For example, for the query information "recommended English courses", a pre-analysis is first performed, and the subdivision dimensions of course content and teaching method are summarized because different learners have different individualization needs for course content, format, etc. Afterwards, based on the historical query information "Beijing online English teaching", "English courses suitable for elementary school students" formed based on the problem, it is analyzed that the supplementary information includes Beijing, online, and elementary school students, and the question dimensions include learning crowd, teaching method, and region. In this way, the question dimensions include {"dimension name": "course content", "options": ["daily", "business", "exam"]}, {"dimension name": "teaching method", "options": ["online", "offline"]}, {"dimension name": "learning crowd", "options": ["elementary school students", "junior high school students", "adults"]}, {"dimension name": "region", "options": ["Beijing", "Shanghai", "Guangzhou"]}.

[0061] In this embodiment, the prior knowledge of the large-scale model can be used to directly analyze the original needs, and provide insight into and suggest the overall needs of the user. Meanwhile, the real and specific needs of the user can be extracted according to the user's search history after the fact, and the reliability of the query content and the applicability of the query can be enhanced. In this way, the subdivision option content can be generated for each query dimension according to the needs. An interactive user selection area can be formed in the search interface, which can guide the user to clarify complex needs.

[0062] In the above-mentioned operation S320, in the process of selecting a target question dimension, a target question dimension can be identified from the multiple question dimensions based on whether the semantic information of the multiple question dimensions matches with the semantic information of the query result related to the query information, and based on the evaluation values ​​of the multiple question dimensions.

[0063] In the above operation S330, in the process of identifying the interaction information, the interaction information can be identified according to the dimension name and multiple options in the target question dimension.

[0064] According to another embodiment of the present disclosure, the process of identifying the plurality of question dimensions may include combining the first presentation information template, the query information, the historical query information, and the query results related to the query information as the first presentation information, and inputting the first presentation information into the large-scale model to obtain the plurality of question dimensions. The first presentation information template may be pre-arranged and stored in a predetermined storage area, and read out from the storage area as needed.

[0065] In one example, the first presented information template includes a first thought chain, the first thought chain includes a plurality of steps, and the first thought chain instructs the large-scale model to reason according to the steps in the first thought chain. For example, the steps in the first thought chain include identifying a relationship type between the query information and the historical query information, identifying whether a question dimension exists based on the relationship type, and for the candidate question dimension, generating question dimensions and options based on a predetermined subdivision direction. In another example, the first presented information template may further indicate the identity and task of the large-scale model. The first presented information template may further indicate the format of the input information and output information of the large-scale model. The first presented information template may further include an example. In this embodiment, the first presented information template includes a first thought chain, and can guide the large-scale model through the steps in the first thought chain to allow the large-scale model to accurately output the question dimension.

[0066] For example, in this embodiment, the query information is "Movies to watch together", and the historical query information includes a plurality of pieces of information, such as videos to watch before two people exercise, movies to watch alone, recommended American movies to watch together, and movies to watch as a couple. Based on the above query information and the historical query information, the following question dimensions can be identified: {'dimension name': 'Movie type', 'options': ['Comedy', 'Action', 'Romance', 'Science Fiction']}, {'dimension name': 'Movie origin', 'options': ['USA', 'China', 'India', 'Other countries']}, {'dimension name': 'Specific attention element', 'options': ['Richness of content', 'Visual effects', 'Word of mouth and evaluation']}.

[0067] For example, in this embodiment, the query information is "top 10 players in mobile game rankings", and the historical query information includes a plurality of pieces of information such as mobile game rankings, top 10 players in martial arts mobile game rankings, top 10 players in mobile game rankings, top 10 players in mobile game rankings, top 10 players in mobile game rankings, etc. Based on the above query information and historical query information, the following question dimensions can be identified: {'dimension name': 'game type', 'options': ['martial arts', 'shooting', 'policy', 'adventure']}, {'dimension name': 'ranking basis', 'options': ['download volume', 'active user volume', 'user usage evaluation']}.

[0068] For example, in this embodiment, the first presentation information template is as follows:

[0069] You are a search query analysis expert and need to ask questions about the "current query" based on the "current query" and "historical query information", refine the user's intent, and guide the user to perform multiple interactive searches. "Refinement" is defined in a specific sense.

[0070] The specific decomposition thinking steps are as follows:

[0071] Step 1, problem and needs analysis: Analyze what information the user is trying to obtain or what problem they are trying to solve through this problem. Determine whether the needs for this problem are clear.

[0072] Step 2, historical query information analysis: Analyze the "historical query information" and judge their relationship with the "current query", where there are mainly four types of relationships: the first type is that the meaning is the same as the "current query" and there is no information gain; the second type is that it is not related to the "current query" at all; the third type is that the subject or need content has changed; and the fourth type is that further subdivision has been made on the "current query". According to the above analysis, judge whether the user historical query information can provide an effective subdivision dimension, and if so, output True, otherwise output False.

[0073] In step 3, it is determined whether the problem needs to be subdivided. If step 2 is not False, the content of the "historical query information" can be referred to in the analysis process. When making a specific determination, it is considered whether the current problem needs to be subdivided from three directions: "subdivision to supplement specific context", "subdivision to match the user's personal situation", and "subdivision of the needs hierarchy". If there is no need to subdivide in any of the three directions, it can be determined that the problem is sufficiently specific and no further subdivision is necessary. Otherwise, it can be determined that the current problem needs to be further subdivided. The problem needs and the "historical query information content" are combined to determine whether the current problem needs to be subdivided.

[0074] In step 4, list the subdivision dimensions: from the three directions of "subdivision that supplements specific context," "subdivision that matches the user's personal situation," and "subdivision of needs hierarchy," list the subdivision dimensions that have a significant impact on the answer to the current problem. Here, the subdivision dimension in the direction of "subdivision of needs hierarchy" does not have any option content, and only the subdivision dimension needs to be provided. The subdivision dimensions in the other two directions need to provide option content.

[0075] Step 5: De-duplicate with existing subdivision dimensions in the current query result: Analyze whether the current query result contains the above decomposed subdivision dimensions, and if so, remove the semantically overlapping dimensions in the next analysis.

[0076] ##Input format: query{}, Answer{}, historical query info{}.

[0077] Output format: { “Problem and Needs Analysis”: “Analyze the user's “current query”. “Historical query information analysis”: “Analyze historical query information behavior.” “Determine whether to subdivide the problem”: “Determine whether to trigger subdivision from three directions.” “List the subdivision dimensions”: “List the subdivision dimensions and corresponding options from three directions.” “Deduplicate with existing subdivision dimensions in the current query results”:“”. }

[0078] ##Example: AAA. ##Identify thought processes and outcomes according to requirements and [examples].

[0079] The first presentation information template has been described above. It will be understood that in the first presentation information template, "query" indicates query information, and "answer" indicates a query result.

[0080] According to another embodiment of the present disclosure, the process of filtering target question dimensions may include a de-duplication sub-process and a value assessment sub-process.

[0081] Regarding the deduplication sub-process, if it is determined that, for each of the multiple question dimensions, the semantic information of the question dimension matches the semantic information of the query result related to the query information, the question dimension can be removed from the multiple question dimensions to obtain a deduplication question dimension.

[0082] For the evaluation sub-process, a target question dimension may be identified from the de-duplicated question dimensions based on the evaluation value of each of the de-duplicated question dimensions.

[0083] In addition, when some contents have already been fed back in the query result, there is no need to repeat the related questions to the user, so that the duplicate queries can be avoided through the duplicate elimination sub-process and the accuracy of the query dimension can be improved.By evaluating the query dimension, the accuracy of the target query dimension can be made to better meet the user query needs.

[0084] According to another embodiment of the present disclosure, the deduplication sub-process in the process of selecting target question dimensions may include combining a second presentation information template, query information, query results and multiple question dimensions as second presentation information, and then inputting the second presentation information into a large-scale model to obtain deduplication-free question dimensions.

[0085] The format of the second presentation information template is similar to that of the first presentation information template. For example, the second presentation information template may be pre-arranged and stored in a predetermined storage area, and read out from the storage area when it is required to be used. For example, the second presentation information template includes a second thought chain, the second thought chain includes a plurality of steps, and the second thought chain instructs the large-scale model to reason according to the steps in the second thought chain. For example, the plurality of steps included in the second thought chain include generating summary information for the query result information, and identifying whether the summary information is related to the question dimension. In another example, the second presentation information template may further indicate the identity and task of the large-scale model. The second presentation information template may further indicate the format of the input information and output information of the large-scale model. The second presentation information template may include an example. In this embodiment, the second presentation information template includes a second thought chain, and by guiding the large-scale model through the steps in the second thought chain, the large-scale model can accurately perform the deduplication process, thereby avoiding asking duplicate questions to the existing content of the query result.

[0086] For example, in this embodiment, the second presentation information template is as follows:

[0087] You are a robot that evaluates whether the question dimensions match the semantics of the query result answers. Your role is to <query> 、 <answer>and <question dimension> to generate a thought logic chain of "whether the current question dimension overlaps with the semantics of the query result".

[0088] The specific thought steps and logic are as follows:

[0089] In step 1, the current <query> 、 <answer>Deeply understand the <Question dimension> and the current situation based on the <Question dimension> <query>Think deeply about and understand the value and impact of the question on your understanding, and summarize and summarise the key points already answered in the current query results.

[0090] In step 2, from the given question dimension list, it is determined whether the content of each question dimension has been mentioned in step 2. If it has been mentioned, it is determined that the question dimension overlaps with the meaning of the current query result; otherwise, there is no overlap.

[0091] In step 3, the corresponding JSON data is output.

[0092] Below we will provide you with examples and you will have to learn how to generate the required thought chain using the examples.

[0093] Example: BBB.

[0094] Next, provide one input and, based on what you have learned above, you will think deeply, understand, and generate a corresponding de-duplication thought chain.

[0095] The second presentation information template has been described above.

[0096] According to another embodiment of the present disclosure, a value assessment sub-process in the process of selecting target question dimensions may include combining a third presentation information template, the query information, the query results and the de-duplicated question dimensions as third presentation information, and then inputting the third presentation information into a large-scale model to obtain target question dimensions.

[0097] The format of the third presentation information template is similar to that of the first presentation information template and the second presentation information template. For example, the third presentation information template may be pre-arranged and stored in a predetermined storage area, and read out from the storage area when it is required for use. For example, the third presentation information template includes a third thought chain, the third thought chain includes a plurality of steps, and the third thought chain instructs the large-scale model to reason according to the steps in the third thought chain. For example, the plurality of steps included in the third thought chain include identifying an evaluation value of a question dimension based on a predetermined evaluation method, and selecting a target question dimension from the deduplicated question dimensions based on the evaluation value, and the predetermined evaluation method may include evaluating the importance of the question dimension to affect the current answer, and may include evaluating from the perspective of whether the question dimension is useful for further understanding the user's needs. In another example, the third presentation information template may further indicate the identity and task of the large-scale model. The third presentation information template may further indicate the format of the input information and output information of the large-scale model. The third presentation information template may further include an example. In this embodiment, the third presentation information template includes a third thought chain, and by guiding the large-scale model through steps in the third thought chain, the large-scale model can accurately select more valuable question dimensions, improve user satisfaction, and ensure the value and applicability of the question dimensions.

[0098] For example, in this embodiment, the third presentation information template is as follows:

[0099] You are a robot to evaluate the question dimension prior value degree. Your role is to <query> 、 <answer>And by using <question dimension>, a thought logic chain of "analyzing the question dimension prior value" is generated.

[0100] The specific thought steps and logic are as follows:

[0101] In step 1, the current <query> 、 <answer>Deeply understand the <Question dimension> and the current situation based on the <Question dimension> <query>Reflect and understand the value and impact of

[0102] In step 2, the value degree of the newly proposed question dimension is evaluated. The specific evaluation criteria are whether the current subdivision direction is an important factor influencing question answering, and whether it is effective in deeply understanding the current question needs.

[0103] Finally, in step 3, the three most valuable target question dimensions are selected from among them, and up to four options are given for each target question dimension, and it is determined whether the options are mutually exclusive.

[0104] Example: CCC.

[0105] The third presentation information template has been described above.

[0106] According to another embodiment of the present disclosure, after obtaining the multiple target question dimensions, the multiple target question dimensions may be sorted, and may be sorted based on the frequency of clicking the target question dimensions. For example, a user group to which the user who inputs the query information belongs may be identified, and then the frequency of clicking the target question dimensions by the user group in the history search process may be identified. Then, the target question dimensions are sorted based on the frequency. Note that the intentions of multiple users in the same user group may be relatively similar, and this embodiment can preferentially present more valuable target question dimensions by identifying the ranking of the target question dimensions according to the history operation of the user group.

[0107] According to another embodiment of the present disclosure, the process of identifying interaction information may include combining the fourth presentation information template, the query information and the target question dimension as fourth presentation information, and then inputting the fourth presentation information into a large-scale model to obtain a question problem in the interaction information.

[0108] The format of the fourth presentation information template is similar to the first presentation information template, the second presentation information template, and the third presentation information template. For example, the fourth presentation information template may be pre-arranged and stored in a predetermined storage area, and read out from the storage area when it is needed to be used. For example, the fourth presentation information template includes a fourth thought chain, the fourth thought chain includes a plurality of steps, and the fourth thought chain instructs the large-scale model to reason according to the steps in the fourth thought chain. For example, the plurality of steps included in the fourth thought chain include identifying a query scene based on the query information and the question dimension, and generating a question problem in the interaction information based on the query scene. In another example, the fourth presentation information template may further indicate the identity and task of the large-scale model. The fourth presentation information template may further indicate the format of the input information and output information of the large-scale model. The fourth presentation information template may further include an example. In this embodiment, the fourth presentation information template includes a fourth thought chain, and guides the large-scale model through the steps in the fourth thought chain, so that the large-scale model can accurately output the question problem.

[0109] For example, in this embodiment, the fourth presentation information template is as follows:

[0110] Your task is to reconstruct this query based on the original search question, the information angle of the query, and the options of the query. Specifically, you need to follow the following steps:

[0111] In step 1, the information angle of the current search problem and query problem is deeply understood, the key factors of the query and the user's needs are analyzed, and the possible background or scene at the time of the user's query is estimated.

[0112] In step 2, based on the user's question background, a simple and fluent question is formed by referring to the information angle and options of the provided question.

[0113] Example: Input: "Search Question": "AA"; "Dimension Name": "BB"; "Choices": "CC". output: “Step 1”: “User question DD, the key factors of this problem are EE and the user needs are FF. The background or scene that the user may encounter is GG, so we need to identify HH”.

[0114] “Step 2”: “The information angle of the query is II, and based on the current query background, a query can be submitted for keyword JJ to guide the user to supplement KK”.

[0115] "Final Question":"LL".

[0116] The fourth presentation information template has been described above.

[0117] In addition, some embodiments of the present disclosure combine a large-scale language model with prompt engineering, and use a presentation information template to guide the large-scale model to output a result. The large-scale model may be a large-scale language model (LLM), a generative model, etc., and the present disclosure does not limit the large-scale model. The presentation information templates (e.g., a pre-processed presentation information template, a first presentation information template, a second presentation information template, a third presentation information template, a fourth presentation information template, and a fifth presentation information template, etc.) may be text information written in a natural language. Each presentation information template may be pre-arranged and stored in a predetermined storage area, and read out from the storage area when it is necessary to use it.

[0118] According to another embodiment of the present disclosure, after the interaction information is obtained, the interaction information can be applied in multiple ways: In one application scenario, the interaction information is displayed through a front-end page, so that the user can make selections according to the query dimension to supplement the query content.

[0119] For example, after a user selects some options, new query information can be newly determined based on the option clicked by the user, and then a query can be made based on the new query information to update the query result displayed to the user. For example, if the query information is "adjusting the body with traditional Chinese medicine", the question dimension name is "gender", and the option selected by the user is "female", the new query information can be "female adjusts the body with traditional Chinese medicine".

[0120] In addition, after the user selects an option, the option selected by the user may be converted into a natural language. For example, the question dimension name may be "gender", the option selected by the user may be "female", and the converted text information may be "My gender is female", and the converted text information may be input into an input box on the page.

[0121] According to another embodiment of the present disclosure, after the interaction information is obtained, the interaction information can be applied in multiple ways. In another application scenario, the interaction information can be used to build a training sample, and the training sample can be used to train another model, and thus, when resources are limited, the other trained model can be used to interact with the user.

[0122] For example, if another model includes two SFT models, for ease of distinction, hereinafter, one SFT model will be referred to as the first model and the other SFT model will be referred to as the second model.

[0123] Regarding the first model, the first model may be a generative model. In the training process, the interaction information may be precisely labeled, and then the first model may be trained using the precisely labeled data as a training sample. After being trained, the input of the first model may include the query information, the historical query information, and the query results related to the query information, and may further include a presentation information template to be applied to the first model. The output of the first model may include a question dimension, a question option, and a question question obtained by natural language conversion of the question dimension.

[0124] For example, the output of the first model is as follows: {"Question dimension":"Name meaning", "Question":"What is the desired meaning of your child's name?", "Options":["Intelligence and wisdom", "Health and peace", "Noble character"], "Are the options mutually exclusive?":"Yes", {"Question dimension":"Name source", "Question":"What is the source of your desired name?", "Options":["Tang Poems", "Song Ci", "Book of Songs", "Chu Ci"], "Are the options mutually exclusive?":"No"}.

[0125] Regarding the second model, the second model may be a generative model. The second model generates a consistent natural language text based on the above question dimensions, question options and question questions, and fills in the input box of the front-end page. For example, the output of the second model is that the gender of my baby is a boy, and I hope to convey the meaning of brave, strong, good and friendly.

[0126] For example, the second model may be guided to generate a natural language query by a pre-arranged presentation information template, such as the following:

[0127] Suppose you are an artificial intelligence assistant, and your main functions are problem analysis and dimension supplementation, and you provide the dimension supplementation to the user in a more natural language expression method, guiding continuous interaction between you and the user.

[0128] The overall analysis process is as follows:

[0129] Thought process: First step, for each dimension, modify the guided description in natural language based on the questioner's perspective in the interaction guide. Second step, for each dimension, randomly select one option and give a need supplement in natural language based on the answerer's perspective in the interaction guide.

[0130] Dimensional Guided Transformation: Based on the first step in the thought process, we give the revised natural language guided description of each dimension.

[0131] Needs Support: Based on the second step in the thought process, we provide a needs support guide.

[0132] Examples are as follows:

[0133] Problem: The difference between a futon and a mattress.

[0134] Supplementary dimensions: [{"dimension name":"contrast aspect", "sub-dimension options":["thickness", "material", "cleanliness"]}].

[0135] Overall analysis process: "Thought process": The supplementary dimension has one dimension called "aspect of contrast", which can be converted into a natural language explanation tailored to the problem, and converted into "When it comes to the distinction between a futon and a mattress, what is the aspect of contrast that interests you more?". Assuming that the user selects "thickness" in the "aspect of contrast" dimension, the supplementary explanation for the user's own needs can be organized, specifically, "I would like to know the difference between a futon and a mattress when the aspect of contrast is thickness."

[0136] "Dimensional Guide Transformation": ["What is the more interesting contrast between a futon and a mattress?"]

[0137] "Additional explanation of needs": {"prefix": "About futons and mattresses": "Combination of dimensional options": ["The contrasting aspect is thickness"], "suffix": "I would like to know the difference between the two"}.

[0138] Note that the output format may be analyzed using a predefined function.

[0139] Based on the above overall analysis process and examples, please analyze the following issues and supplementary dimensions.

[0140] Problem: MMM Supplementary dimensions: NNN

[0141] The presentation information template that guides the second model has been described above.

[0142] FIG. 4 is a schematic configuration block diagram of an apparatus for identifying interaction information according to an embodiment of the present disclosure.

[0143] As shown in FIG. 4 , the apparatus 400 for identifying interaction information may include a first identifying module 410 , a second identifying module 420 , and a third identifying module 430 .

[0144] The first identification module 410 identifies a plurality of question dimensions based on the target query information and the historical query information, where each question dimension includes a dimension name and a plurality of options.

[0145] The second identification module 420 identifies a target question dimension from the multiple question dimensions based on whether the semantic information of the multiple question dimensions matches the semantic information of the query result related to the query information, and based on the evaluation values ​​of the multiple question dimensions.

[0146] The third identification module 430 identifies the interaction information based on the dimension names and the multiple options in the target question dimension.

[0147] In another embodiment of the present disclosure, the above-mentioned device further includes a history information identification module that identifies, among the candidate history query information, candidate history query information that satisfies the following conditions: an intent of the candidate history query information matches an intent of the query information, a search frequency of the candidate history query information satisfies a predetermined frequency condition, and an entity included in the candidate history query information matches an entity included in the query information, as the history query information.

[0148] In another embodiment of the present disclosure, the first identification module includes a first combination submodule and a first input submodule, the first combination submodule combines the first presentation information template, the query information, the historical query information, and the query result related to the query information as the first presentation information, the first input submodule inputs the first presentation information into the large-scale model to obtain a plurality of question dimensions, the first presentation information template includes a first thought chain, the first thought chain instructs the large-scale model to obtain a plurality of question dimensions in the following manner: identify a relationship type between the query information and the historical query information, identify whether a question dimension exists according to the relationship type, and generate a question question and a choice according to a predetermined subdivision direction for a candidate question dimension.

[0149] In another embodiment of the present disclosure, the second identifying module includes a deleting submodule and a identifying submodule, where when the deleting submodule identifies, for each of the plurality of question dimensions, that the semantic information of the question dimension matches the semantic information of the query result related to the query information, the deleting submodule deletes the question dimension from the plurality of question dimensions to obtain a de-duplicated question dimension, and the identifying submodule identifies a target question dimension from the de-duplicated question dimensions based on the evaluation value of each of the de-duplicated question dimensions.

[0150] In another embodiment of the present disclosure, the deletion submodule includes a first combination unit and a first input unit, where the first combination unit combines a second presentation information template, query information, a query result and a plurality of question dimensions as a second presentation information, and the first input unit inputs the second presentation information into a large-scale model to obtain a de-duplication question dimension, where the second presentation information template includes a second thought chain, and the second thought chain instructs the large-scale model to obtain a de-duplication question dimension in the following manner: generate summary information for the query result, and determine whether the summary information is related to the question dimension.

[0151] In another embodiment of the present disclosure, the specific submodule includes a second combination unit and a second input unit, the second combination unit combines a third presentation information template, the query information, the query result, and the de-duplicated question dimension as a third presentation information, and the second input unit inputs the third presentation information into the large-scale model to obtain a target question dimension, where the third presentation information template includes a third thought chain, and the third thought chain instructs the large-scale model to obtain the target question dimension in the following manner: determine an evaluation value of the question dimension according to a predetermined evaluation manner, and select a target question dimension from the de-duplicated question dimensions according to the evaluation value.

[0152] In another embodiment of the present disclosure, the third identification module includes a second combination submodule and a second input submodule, the second combination submodule combines a fourth presentation information template, query information and a target question dimension as a fourth presentation information, the second input submodule inputs the fourth presentation information into the large-scale model to obtain a question problem in the interaction information, where the fourth presentation information template includes a fourth thought chain, and the fourth thought chain instructs the large-scale model to obtain a question problem in the interaction information in the following manner: identify a query scene according to the query information and the question dimension, and generate a question problem in the interaction information according to the query scene.

[0153] According to an embodiment of the present disclosure, the present disclosure further provides an electronic device including at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions, when executed by the at least one processor, enable the at least one processor to perform the method of identifying interaction information.

[0154] According to an embodiment of the present disclosure, the present disclosure further provides a non-transitory computer-readable storage medium having computer instructions stored thereon, where the computer instructions cause a computer to perform the method of identifying interaction information as described above.

[0155] According to an embodiment of the present disclosure, the present disclosure further provides a computer program product that, when executed by a processor, realizes the method for identifying interaction information.

[0156] 5 is a structural block diagram of an electronic device for implementing the method for identifying interaction information according to an embodiment of the present disclosure. The electronic device is intended to represent various types of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. The electronic device may also represent various types of mobile devices, such as personal digital assistants, mobile phones, smartphones, wearable devices, and other similar computing devices. The components, their connections and relationships, and their functions shown herein are merely exemplary and do not limit the implementation of the present disclosure as described and / or claimed herein.

[0157] 5, the device 500 includes a computing unit 501, which may perform various appropriate operations and processes based on a computer program stored in a read-only memory (ROM) 502 or loaded from a storage unit 508 into a random access memory (RAM) 503. The RAM 503 may further store various programs and data required for the operation of the device 500. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0158] The components of the device 500 are connected to an I / O interface 505, which includes an input unit 506, such as a keyboard, a mouse, etc., an output unit 507, such as various types of displays, speakers, etc., a storage unit 508, such as a magnetic disk, an optical disk, etc., and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 enables the device 500 to exchange information and data with other devices via a computer network such as the Internet and / or various electrical networks.

[0159] The computing unit 501 may be various general-purpose and / or dedicated processing modules having processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, computing units running various machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 501 performs each of the methods and processes described above, such as the method of determining interaction information. For example, in some embodiments, the method of determining interaction information may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 508. In some embodiments, some or all of the computer program may be loaded and / or installed in the device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, it may perform one or more steps of the method of determining interaction information described above. Alternatively, in other embodiments, the computing unit 501 may be configured to perform the method of determining interaction information in any other suitable manner (eg, via firmware).

[0160] Various embodiments of the systems and techniques described herein above may be realized in digital electronic circuitry systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may be embodied in one or more computer programs that may be executed and / or interpreted by a programmable system that includes at least one programmable processor, which may be a special purpose or general purpose programmable processor, and may include capable of receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.

[0161] The program codes for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, a special purpose computer or other programmable data processing apparatus, so that when the program codes are executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are performed. The program codes may be fully executed on the device, partially executed on the device, partially executed on the device as a separate software package and partially executed on a remote device or fully executed on a remote device or server.

[0162] In the context of this disclosure, a machine-readable medium may be a tangible medium, and may contain or store a program for use in or in combination with an instruction execution system, device, or electronic device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or any suitable combination of the above. More specific examples of machine-readable storage media include electrical connections through one or more wires, portable computer disks, hard disks, random access memories (RAMs), read-only memories (ROMs), erasable programmable read-only memories (EPROMs or flash memories), optical fibers, compact disk read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0163] A computer may implement the systems and techniques described herein to provide for interaction with a user, the computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user, and a keyboard and pointing device (e.g., a mouse or trackball) through which a user can provide input to the computer. Other types of devices may also provide for interaction with a user, for example, the feedback provided to the user may be any form of sensing feedback (e.g., visual feedback, auditory feedback, or tactile feedback) and may receive input from the user in any form (including voice input, speech input, or tactile input).

[0164] The systems and techniques described herein may be implemented in a computing system including background components (e.g., a data server), or a computing system including middleware components (e.g., an application server), or a computing system including front-end components (e.g., a user computer having a graphical user interface or a web browser through which a user may interact with embodiments of the systems and techniques described herein), or a computing system including any combination of such background, middleware, or front-end components. The components of the system may be connected to each other by any form or medium of digital data communication (e.g., a communications network). Examples of communications networks illustratively include a local area network (LAN), a wide area network (WAN), and the Internet.

[0165] A computer system may include clients and servers. Clients and servers are generally remote and typically interact through a communication network. The relationship of client and server is created by computer programs running on the corresponding computers and having a client-server relationship.

[0166] It should be understood that various types of flows shown above may be used, and operations may be rearranged, added, or removed. For example, each operation described in the present disclosure may be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and the present specification is not limited thereto.

[0167] The above specific embodiments do not limit the scope of protection of the present disclosure. Those skilled in the art should understand that various modifications, combinations, subcombinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principle of the present disclosure should be included within the scope of protection of the present disclosure.< / query> < / answer> < / query> < / answer> < / query> < / query> < / answer> < / query> < / answer> < / query>

Claims

1. 1. A method for identifying interaction information, comprising: identifying a plurality of query dimensions based on the target query information and the historical query information, each query dimension including a dimension name and a plurality of options; Identifying a target question dimension from the plurality of question dimensions based on whether semantic information of the plurality of question dimensions matches semantic information of a query result related to the query information and based on evaluation values ​​of the plurality of question dimensions; and identifying interaction information based on a dimension name and a plurality of options in the target question dimension. A method for identifying interaction information, comprising:

2. The method further includes identifying, as the historical query information, candidate history query information that satisfies the following conditions: an intent of the candidate history query information matches an intent of the query information, a search frequency of the candidate history query information satisfies a predetermined frequency condition, and an entity included in the candidate history query information matches an entity included in the query information.

2. The method of claim 1 .

3. Identifying a plurality of query dimensions based on the subject query information and the historical query information includes: combining a first presentation information template, the query information, the historical query information, and query results related to the query information as first presentation information; inputting the first submitted information into a large-scale model to obtain the plurality of question dimensions; Wherein, the first presentation information template includes a first thought chain, and the first thought chain instructs the large-scale model to obtain the plurality of question dimensions in the following manner: identify a relationship type between the query information and the historical query information; identify whether a question dimension exists based on the relationship type; and, for a candidate question dimension, generate a question dimension and an option based on a predetermined subdivision direction.

2. The method of claim 1 .

4. Identifying a target question dimension from the plurality of question dimensions based on whether semantic information of the plurality of question dimensions matches semantic information of a query result related to the query information, and based on evaluation values ​​of the plurality of question dimensions, When it is determined that semantic information of each of the plurality of question dimensions matches semantic information of the question dimension and semantic information of a query result related to the query information, the question dimension is deleted from the plurality of question dimensions to obtain a question dimension after deduplication; and identifying the target query dimension from the query dimensions after the duplicates are eliminated based on the evaluation values ​​of the query dimensions after the duplicates are eliminated.

2. The method of claim 1 .

5. When it is determined that semantic information of the question dimension matches semantic information of a query result related to the query information, removing the question dimension from the plurality of question dimensions to obtain a question dimension after de-duplication is eliminated, combining a second presentation information template, the query information, the query results, and the plurality of question dimensions as a second presentation information; inputting the second submitted information into a large-scale model to obtain a question dimension after de-duplication; Here, the second presentation information template includes a second thought chain, and the second thought chain instructs the large-scale model to obtain a de-duplicated question dimension by generating summary information for the query result and determining whether the summary information is related to the query dimension.

5. The method of claim 4.

6. Identifying a target query dimension from the query dimensions after the duplicate elimination based on the evaluation value of each of the query dimensions after the duplicate elimination, combining a third presentation information template, the query information, the query result, and the de-duplicated question dimensions as a third presentation information; inputting the third submitted information into a large-scale model to obtain the target question dimension; Here, the third presentation information template includes a third thought chain, and the third thought chain instructs the large-scale model to obtain the target question dimension by identifying an evaluation value of a question dimension based on a predetermined evaluation method, and selecting the target question dimension from the question dimensions after the duplicates are eliminated based on the evaluation value.

5. The method of claim 4.

7. Identifying interaction information based on a dimension name and a plurality of options in the target question dimension, combining a fourth presentation information template, the query information, and the target question dimension as a fourth presentation information; inputting the fourth presentation information into a large-scale model to obtain a question in the interaction information; Wherein, the fourth presentation information template includes a fourth thought chain, and the fourth thought chain instructs the large-scale model to obtain a question question in the interaction information in the following manner: identifying a query scene according to the query information and the question dimension; and generating a question question in the interaction information according to the query scene.

2. The method of claim 1 .

8. 1. An apparatus for identifying interaction information, comprising: a first identification module for identifying a plurality of query dimensions based on the target query information and the historical query information, each query dimension including a dimension name and a plurality of options; a second identification module that identifies a target question dimension from the plurality of question dimensions based on whether semantic information of the plurality of question dimensions matches semantic information of a query result related to the query information and based on an evaluation value of the plurality of question dimensions; and a third identification module for identifying interaction information based on a dimension name and a plurality of options in the target question dimension. An apparatus for identifying interaction information, comprising:

9. The method further includes a history information identification module that identifies, as the history query information, candidate history query information that satisfies the following conditions: an intent of the candidate history query information matches an intent of the query information, a search frequency of the candidate history query information satisfies a predetermined frequency condition, and an entity included in the candidate history query information matches an entity included in the query information.

9. The apparatus of claim 8.

10. The first identification module is a first combination submodule that combines a first presentation information template, the query information, the historical query information, and a query result related to the query information as a first presentation information; a first input submodule for inputting the first submitted information into a large-scale model to obtain the plurality of question dimensions; Wherein, the first presentation information template includes a first thought chain, and the first thought chain instructs the large-scale model to obtain the plurality of question dimensions in the following manner: identify a relationship type between the query information and the historical query information; identify whether a question dimension exists based on the relationship type; and, for a candidate question dimension, generate a question dimension and an option based on a predetermined subdivision direction.

9. The apparatus of claim 8.

11. The second identification module is a deletion submodule for deleting the question dimension from the plurality of question dimensions when it is determined that the semantic information of the question dimension matches the semantic information of the query result related to the query information, and acquiring a question dimension after deduplication; and a specifying submodule for specifying the target question dimension from the post-duplicate elimination question dimensions based on the evaluation value of each of the post-duplicate elimination question dimensions. Apparatus according to any one of claims 8 to 10.

12. The deletion submodule: a first combination unit for combining a second presentation information template, the query information, the query result, and the plurality of question dimensions as a second presentation information; a first input unit for inputting the second submitted information into a large-scale model to obtain a question dimension after redundancy elimination; Here, the second presentation information template includes a second thought chain, and the second thought chain instructs the large-scale model to obtain a de-duplicated question dimension by generating summary information for the query result and determining whether the summary information is related to the question dimension.

12. The apparatus of claim 11 .

13. The specific submodule is a second combination unit for combining a third presentation information template, the query information, the query result, and the question dimension after the duplicates are eliminated as a third presentation information; a second input unit for inputting the third submitted information into a large-scale model to obtain the target question dimension; Here, the third presentation information template includes a third thought chain, and the third thought chain instructs the large-scale model to obtain the target question dimension by identifying an evaluation value of a question dimension based on a predetermined evaluation method, and selecting the target question dimension from the question dimensions after the duplicates are eliminated based on the evaluation value.

12. The apparatus of claim 11 .

14. The third identification module is a second combination submodule for combining a fourth presentation information template, the query information, and the target question dimension as a fourth presentation information; a second input submodule for inputting the fourth presentation information into a large-scale model and obtaining a question in the interaction information; Wherein, the fourth presentation information template includes a fourth thought chain, and the fourth thought chain instructs the large-scale model to obtain a question question in the interaction information in the following manner: identifying a query scene according to the query information and the question dimension; and generating a question question in the interaction information according to the query scene.

9. The apparatus of claim 8.

15. At least one processor; a memory in communication with the at least one processor; The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor such that the at least one processor can perform the method according to any one of claims 1 to 7. electronic equipment.

16. A non-transitory computer-readable storage medium having computer instructions stored thereon, comprising: The computer instructions cause a computer to carry out the method according to any one of claims 1 to 7. A non-transitory computer-readable storage medium.

17. A computer program which, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images