Information feedback method and device, electronic equipment and storage medium

Abstract

The invention relates to an information feedback method and device, electronic equipment and a storage medium. The method comprises the following steps: in response to received input information of a user, obtaining a query result from a target database based on the input information; based on the query result, processing the input information into cue word data; based on the prompt word data, feedback information for the user is generated through a large model; wherein the target database comprises a database which is dynamically updated according to a new word determination result and / or a hot word determination result. According to the scheme provided by the invention, the target database can update the content in the target database in time by dynamically updating the database according to the new word determination result and / or the hot word determination result, so that the timeliness of the content in the target database is relatively high, the content lag in the target database is avoided, and the user experience is improved. Therefore, more accurate knowledge which is easier to meet user expectation is provided for the large model, accurate feedback information can be output to the user, and the fluency of communication with the user and the user experience are improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to an information feedback method, device, electronic device and storage medium. Background Art

[0002] In related technologies, with the development of artificial intelligence technology, intelligent systems based on retrieval-augmented generation (RAG) have been widely used in fields such as information retrieval and content generation.

[0003] However, with the development of the Internet, new words, hot words and buzzwords are constantly emerging, and traditional speech recognition systems have difficulty in timely identifying these new words. This has led to a decline in user experience in the actual use of artificial intelligence technology. Summary of the Invention

[0004] In order to solve or partially solve the problems existing in the related art, the present application provides an information feedback method, device, electronic device and storage medium. By dynamically updating the database based on the new word determination results and / or hot word determination results, the target database can update the content in the target database in a timely manner, so that the content in the target database is more timely, avoiding the lag of the content in the target database, thereby providing the large model with more accurate knowledge that is easier to meet user expectations, which is conducive to outputting accurate feedback information to the user, improving the fluency of communication with the user and the user experience.

[0005] The first aspect of the present application provides an information feedback method, comprising: in response to receiving user input information, obtaining a query result from a target database based on the input information; based on the query result, processing the input information into prompt word data; based on the prompt word data, generating feedback information for the user through a large model; wherein the target database includes a database that is dynamically updated according to new word determination results and / or hot word determination results.

[0006] In some embodiments, processing the input information into prompt word data based on the query result includes: generating reference information based on the query result; and processing the input information into prompt word data using the reference information.

[0007] In some embodiments, the query result includes a first entity word and a second entity word having a corresponding relationship; generating reference information based on the query result includes: inserting the first entity word and the second entity word into preset slots of the reference information in a one-to-one correspondence to generate the reference information.

[0008] In some embodiments, feedback information for the user is generated through a large model based on the prompt word data, including: determining the user's intention based on the element content of the input information; and generating feedback information for the user through a large model based on the prompt word data and the user's intention.

[0009] In some embodiments, the input information is current round input information; obtaining query results from the target database based on the input information includes: obtaining historical conversation information; and obtaining query results from the target database based on the current round input information and the historical conversation information.

[0010] In some embodiments, the method further includes: in response to receiving a timeliness data change task processing list, determining a new word determination result and / or a hot word determination result according to the timeliness data change task processing list; and updating the entity words in the target database according to the new word determination result and / or the hot word determination result.

[0011] In some embodiments, the method further includes: obtaining target data through real-time data synchronization processing; based on the target data, obtaining new word determination results and / or hot word determination results using a hot word calculation algorithm and / or a new word calculation algorithm; and generating a time-sensitive data change task processing list based on the new word determination results and / or the hot word determination results.

[0012] The second aspect of the present application provides an information feedback device, which includes: an acquisition module for obtaining a query result from a target database based on the input information in response to receiving the user's input information; a processing module for processing the input information into prompt word data based on the query result; a generation module for generating feedback information for the user through a large model based on the prompt word data; wherein the target database includes a database that is dynamically updated according to the new word determination results and / or hot word determination results.

[0013] In some embodiments, the device also includes: a first update module, used to determine the new word determination result and / or the hot word determination result according to the timeliness data change task processing list in response to receiving the timeliness data change task processing list; a second update module, used to update the entity words in the target database according to the new word determination result and / or the hot word determination result.

[0014] A third aspect of the present application provides an electronic device, including: processor; and The memory stores executable codes thereon, and when the executable codes are executed by the processor, the processor is caused to execute the method described above.

[0015] A fourth aspect of the present application provides a computer-readable storage medium having executable code stored thereon. When the executable code is executed by a processor of an electronic device, the processor is caused to execute the method described above.

[0016] The technical solution provided by this application may have the following beneficial effects: The technical solution provided by this application can enable the target database to update the content in the target database in a timely manner by dynamically updating the database based on the new word determination results and / or hot word determination results, so that the content in the target database is more timely, avoiding the lag of content in the target database, thereby providing the large model with more accurate knowledge that is easier to meet user expectations, which is conducive to outputting accurate feedback information to users, improving the fluency of communication with users and user experience.

[0017] The technical solution of the present application can also: analyze user intentions, provide feedback based on user intentions, and improve feedback efficiency and user satisfaction.

[0018] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The above and other objects, features and advantages of the present application will become more apparent by describing in more detail the exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.

[0020] Figure 1 1 is a flow chart of an information feedback method according to an embodiment of the present application; Figure 2 is another flow chart of the information feedback method shown in an embodiment of the present application; Figure 3 This is a flow chart of intelligent voice interaction according to an embodiment of the present application; Figure 4 is a structural diagram of an information feedback device shown in an embodiment of the present application; Figure 5 It is a structural diagram of an electronic device shown in an embodiment of the present application. DETAILED DESCRIPTION

[0021] The following describes embodiments of the present application in more detail with reference to the accompanying drawings. Although the accompanying drawings illustrate embodiments of the present application, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0022] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. As used in this application and the appended claims, the singular forms "a," "an," "the," and "the" are intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0023] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0024] Large Language Models (LLMs), as a large model, have attracted considerable attention for their exceptional linguistic capabilities and demonstrated intelligence. They can fulfill practical daily needs, such as document summarization and copywriting generation, and can also answer knowledge-based questions.

[0025] However, large language model systems have shortcomings in handling the timeliness and dynamic nature of knowledge across multiple rounds of interaction. When processing dynamic knowledge, especially information with high timeliness (such as artist song titles, location points of interest, etc.), the knowledge base lags due to its real-time nature and rapid changes. Large language model systems are unable to acquire and update this highly time-sensitive information in real time. Waiting for updates from the underlying large model or relying on its generalization capabilities results in a low success rate, resulting in inaccurate information for users during interactions and impacting the smoothness of voice interaction and user experience.

[0026] In response to the above problems, an embodiment of the present application provides an information feedback method, which enables the target database to update the content in the target database in a timely manner by dynamically updating the database based on the new word determination results and / or hot word determination results, so that the content in the target database is more timely, avoiding the lag of content in the target database, thereby providing the large model with more accurate knowledge that is easier to meet user expectations, which is conducive to outputting accurate feedback information to users, improving the fluency of communication with users and user experience.

[0027] The technical solutions of the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0028] Figure 1 It is a flow chart of the information feedback method shown in an embodiment of the present application.

[0029] See also Figure 1 , an information feedback method, the method comprising: Step 101: In response to receiving user input information, obtain a query result from a target database based on the input information.

[0030] It is understood that the target database in this step can be a database that is dynamically updated based on the new word determination results and / or hot word determination results. For example, feedback control technology can be used to maintain the timeliness and data freshness of the content in the target database in a dynamic environment.

[0031] It can be understood that new words refer to words that have emerged with the development of the times or old words that have been used in new ways. Proper nouns in specific fields can also be regarded as new words.

[0032] It's understandable that buzzwords are phrases that are summarized from hot events and have a specific event-specific meaning. These words are often created by netizens and reflect the core information and public sentiment of social public events. Buzzwords are not only a reflection of linguistic development but also a reflection of youth self-awareness and social change.

[0033] Step 102: Based on the query result, the input information is processed into prompt word data.

[0034] It can be understood that prompt word data is key information used to guide the large model to generate specific types of answers or content. They are like a key that can open the "door of thinking" of the large model, allowing it to think and answer questions according to the user's wishes.

[0035] Through carefully designed prompt words, the accuracy and efficiency of large models in practical applications can be improved.

[0036] Step 103: Generate feedback information for the user through the large model based on the prompt word data.

[0037] A large model refers to a large language model, which is a machine learning model that can understand and generate human language.

[0038] Feedback information refers to the feedback answer generated based on the prompt word data.

[0039] The information feedback method provided in this embodiment can be applied to large model systems and multi-round dialogue scenarios, such as task scenarios where users conduct intelligent voice interaction with the cockpit.

[0040] After receiving the user's input information, a target database with high timeliness can be called, and the content stored in the target database can be obtained from the target database through the input information to obtain the query result.

[0041] In a multi-round dialogue scenario, the input information may include the current round input information and historical dialogue information. In some embodiments, the input information may also be the current round input information.

[0042] In some embodiments, obtaining a query result from a target database based on input information may include: obtaining historical conversation information; and obtaining a query result from the target database based on the current round of input information and the historical conversation information.

[0043] Based on the query results, the input information is processed to obtain prompt word data that can guide the large model to generate accurate answers.

[0044] In some embodiments, processing the input information into prompt word data based on the query result may include: generating reference information based on the query result; and processing the input information into prompt word data using the reference information.

[0045] For example, the target database maintains a vocabulary, such as a singer-song-work relationship vocabulary. Therefore, the query results may include a first entity word (e.g., singer) and a second entity word (e.g., song) that have a corresponding relationship. Therefore, the query result may be "X singer - X song," which has a first entity word and a second entity word that have a corresponding relationship.

[0046] Furthermore, generating reference information based on the query result may include: inserting the first entity word and the second entity word into preset slots of the reference information in a one-to-one correspondence to generate the reference information.

[0047] It can be understood that when there is content in the input information that matches the query word list in the target database, the corresponding query results can be obtained, and then reference information can be constructed through the query results, such as filling an entity word and a second entity word (a singer-a song) into the reference information slot one by one to complete the construction, thereby generating the reference information, such as the singer of a certain song is a certain singer.

[0048] Furthermore, the reference information "the singer of a certain song is a certain singer" is used to process the input information and generate feedback information for the user.

[0049] The information feedback method of the embodiment of the present application can enable the target database to update the content in the target database in a timely manner by dynamically updating the database based on the new word determination results and / or hot word determination results, so that the content in the target database is more timely, avoiding the lag of content in the target database, thereby providing the large model with more accurate knowledge that is easier to meet user expectations, and thus facilitating the output of accurate feedback information to the user, improving the fluency of communication with the user and the user experience.

[0050] In some embodiments, feedback information for the user is generated through a large model based on the prompt word data, including: determining the user's intention based on the element content of the input information; and generating feedback information for the user through a large model based on the prompt word data and the user's intention.

[0051] It can be understood that the element content is part of the input information.

[0052] Element information refers to the information content in the input information that can represent the user's dialogue intention, or the information content related to the user's dialogue intention.

[0053] In this embodiment, based on the prompt word data and user intention, feedback information for the user is generated through a large model. It is possible to first determine whether the prompt word data and the user intention match. For example, if the prompt word data indicates that the user wants to listen to a song by a certain singer, and the user intention is to play it, then they are a match, and feedback information for the user is generated, such as an instruction: play a song by a certain singer.

[0054] For example, if the prompt word data indicates that the user wants to listen to a song by a certain singer, but the user's intention is not to play but to chat, then it is a mismatch. In this case, feedback information is generated for the user, such as an instruction: A song sung by a certain singer is really nice.

[0055] The information feedback method of the embodiment of the present application generates feedback information for the user through a large model based on prompt word data and user intention, can analyze user intention, give feedback based on user intention, and improve feedback efficiency and user satisfaction.

[0056] Figure 2 This is another flow chart of the information feedback method shown in an embodiment of the present application.

[0057] See also Figure 2 , the information feedback method also includes: Step 201: Obtain target data through real-time data synchronization processing.

[0058] Step 202: Based on the target data, a hot word calculation algorithm and / or a new word calculation algorithm are used to obtain a new word determination result and / or a hot word determination result.

[0059] Step 203: Generate a time-sensitive data change task processing list based on the new word determination results and / or hot word determination results.

[0060] Step 204 : In response to receiving the timeliness data change task processing list, determining a new word determination result and / or a hot word determination result according to the timeliness data change task processing list.

[0061] Step 205: Update entity words in the target database according to the new word determination result and / or the hot word determination result.

[0062] It can be understood that real-time data synchronization is an important step in the dynamic update of the database. Real-time data synchronization can be achieved through data replication, message queues and stream processing frameworks.

[0063] Hot word calculation algorithms are used to identify words that have rapidly grown and attracted attention within a specific time period, thereby determining hot word results. By analyzing changes in word frequency, we can extract keywords with significant popularity. Examples of hot word ranking calculation methods include the Bayesian average method and Newton's law of cooling.

[0064] The new word calculation algorithm can establish a rule library, professional vocabulary library or pattern library based on the word-forming characteristics or appearance characteristics of new words. By writing regular expressions, these contents can be extracted from the article, and cleaned and filtered according to the corresponding rules. Then, new words are discovered through rule matching to obtain the new word determination results.

[0065] It is understood that the aforementioned hot word determination results and / or new word determination results can be added to an inactive "Time-Effective Data Change Task Processing List." When the "Time-Effective Data Change Task Processing List" is triggered to take effect, if the timeliness status becomes current, the "Time-Effective Data Change Task Processing List" is sent. Thus, in response to receiving the Time-Effective Data Change Task Processing List, the new word determination results and / or hot word determination results are determined based on the Time-Effective Data Change Task Processing List.

[0066] Furthermore, the entity words in the target database may be updated according to the new word determination results and / or the hot word determination results, so as to prevent the target database from being out of knowledge.

[0067] The information feedback method of the embodiment of the present application determines the new word determination result and / or the hot word determination result according to the timeliness data change task processing list in response to receiving the timeliness data change task processing list; updates the entity words in the target database according to the new word determination result and / or the hot word determination result, so that the target database can update the content in the target database in a timely manner, so that the content in the target database is more timely, avoiding the lag of content in the target database.

[0068] In order to better understand the present application, the content of the present application is further described below in conjunction with embodiments, such as taking a dynamic knowledge base of singers' songs as an example. However, the present application is not limited to the following embodiments.

[0069] For example, in the scenario where users conduct intelligent voice interaction with the cockpit, the large models in related technologies are insufficient in processing the timeliness and dynamics of knowledge in multiple rounds of interactions.

[0070] For example, taking playing a singer and his songs as an example, the user said in the previous round, history: play singer A's songs. In this round, he said, query: play song A.

[0071] Then the NLU model knows that "Song A" is sung by singer A, so it rewrites the query in this round as: play song A by singer A. However, if a singer recently sang a new song and the model does not include this knowledge, it is easy to misunderstand the user's intention.

[0072] For example, if a user said, "History: Play songs by singer B," in the previous round, and then said, "Query: Play songs by singer B" in this round, the model would assume that the user wanted to play singer A's classic song "Song B." However, considering that singer B recently covered "Song B," the user's intention is more likely to be singer B's cover of "Song B."

[0073] This problem is caused by the fact that new words / hot words in the open domain cannot be updated as dynamic knowledge. Therefore, it is necessary to optimize the current solution to deal with the lack of timeliness and dynamism of knowledge in multiple rounds of inheritance interactions.

[0074] In order to improve the adaptability and updating efficiency of dynamic knowledge, the dynamic knowledge of the cockpit NLU can be quickly detected and updated to ensure that the voice system can be adjusted in time when new words / hot words in the open domain change, thereby improving the accuracy of command recognition and execution.

[0075] For example, it can enhance the dynamic management capabilities of information with strong timeliness (such as singer song titles, location POI names, etc.), enabling the system to obtain and update this dynamic knowledge in real time, enabling the system to flexibly adapt to system upgrades and changes in user needs, ensuring that users obtain accurate and up-to-date information when querying or operating, and improving the fluency of voice interaction and user experience.

[0076] For example, in the current natural language understanding (NLU) large model system, adding a real-time updated singer-song work relationship vocabulary can solve the problem of untimely knowledge updates.

[0077] For example, supervised fine-tuning of the SFT allows the large model to learn the rules and methods for using reference knowledge, allowing the necessary reference information about the artist-song relationship to be input during the large model's inference. During multi-round conversations, the NLU large model system queries the artist-song relationship vocabulary, identifies the relationship between artists and songs in historical conversations, and provides this information to the large model through retrieval-enhanced generation, helping the large model decide whether to supplement historical artist information for the current round of conversation. Considering whether the artist in the previous round has performed with the current song, and whether the user's intention is to play the song, three scenarios can be categorized, as shown in Table 1: 1. If there is singer and song information in the historical conversation and a singing relationship exists, and the user intends to play the song, the singer mentioned in the historical information will be completed in this round.

[0078] 2. If there is no singing relationship, this round will not be completed.

[0079] 3. If there is a singer-song relationship, but the user's intention is not to play the song, the reference information will be ignored and no completion will be performed.

[0080] Table 1

[0081] For example, for dynamic knowledge such as singer-song relationships, a local vocabulary table can be maintained; for example, the key is the singer's name and the value is the song they sang. Design a large model prompt, and include instructions for using external reference knowledge in the prompt instructions and thinking process. Then, using this prompt, add a few-shot learning example, access the NLU large model system of the current embodiment, obtain failure cases, and generate the training and test sets required for the natural language processing model to obtain a trained target model.

[0082] Figure 3 It is a flow chart of intelligent voice interaction shown in an embodiment of the present application.

[0083] See also Figure 3The NLU large-scale model system acquires interaction information from multiple rounds of interaction, namely historical conversation information 301 and current-round input information 302. For example, historical conversation information 301 includes query-1: "Open the car window" and query-2: "Play a song by singer C." Current-round input information 302 includes query-n: "Play song C." The interaction information is sent to a database 303 for a vocabulary query, resulting in query results 304, such as "Singer: singer C" and "Song: song C." Reference information is then constructed, with entity words (singer and song) correspondingly assigned to the reference information slots. This construction completes the process, resulting in reference information 305, such as "The singer of song C is singer C." If no singing relationship exists, the reference information is empty, and no dynamic knowledge is injected. The reference information is then injected into the prompt as dynamic knowledge, for example, by sending reference information 305. Reference information 305 is then used to process the interaction information (historical conversation information 301 and current-round input information 302) into prompt word data 306, such as "The user wants to listen to song C by singer C." Finally, the prompt word data 306 is input into the large model 307, and the large model 307 is assisted in making a decision and outputting feedback information 308, such as playing song C by singer C.

[0084] This embodiment of the application addresses dynamic, time-sensitive knowledge (such as new songs by artists and popular locations) by maintaining a locally updated, real-time singer-song relationship vocabulary and training a large model through supervised fine-tuning to learn the rules and methods for using reference knowledge. This solves the problem in related technologies where large models, due to lagging knowledge bases and the inability to update in real time, lead to inaccurate information.

[0085] The embodiment of the present application can be applied to multi-round dialogue scenarios. The locally maintained singer-song relationship vocabulary is queried through the NLU large model system to identify the relationship between singers and songs in historical dialogues, and the reference information is provided to the large model to assist in decision-making, thereby improving the flexibility of the voice system and enhancing the fluency of voice interaction and user experience.

[0086] Corresponding to the aforementioned application function implementation method embodiment, the present application also provides an information feedback device, an electronic device and corresponding embodiments.

[0087] Figure 4 It is a structural diagram of the information feedback device shown in an embodiment of the present application.

[0088] See also Figure 4 The information feedback device 400 of this embodiment includes an obtaining module 410 , a processing module 420 and a generating module 430 .

[0089] The obtaining module 410 is configured to obtain a query result from a target database based on the input information received from the user.

[0090] The processing module 420 is used to process the input information into prompt word data based on the query result.

[0091] The generation module 430 is used to generate feedback information for the user based on the prompt word data through the large model.

[0092] Regarding the apparatus in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated again here.

[0093] According to embodiments of the present application, any multiple modules among the acquisition module 410, processing module 420, and generation module 430 may be combined into a single module, or any one of these modules may be split into multiple modules. Alternatively, at least part of the functionality of one or more of these modules may be combined with at least part of the functionality of other modules and implemented in a single module. According to embodiments of the present disclosure, at least one of the acquisition module 410, processing module 420, and generation module 430 may be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or may be implemented in hardware or firmware through any other reasonable means of circuit integration or packaging, or may be implemented in any one of the three implementation methods of software, hardware, and firmware, or any appropriate combination of any of these. Alternatively, at least one of the acquisition module 410, processing module 420, and generation module 430 may be at least partially implemented as a computer program module that, when executed, performs the corresponding functionality.

[0094] Figure 5 It is a structural diagram of an electronic device shown in an embodiment of the present application.

[0095] See also Figure 5 , the electronic device 500 includes a memory 510 and a processor 520.

[0096] The processor 520 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0097] Memory 510 may include various types of storage units, such as system memory, read-only memory (ROM), and permanent storage. ROM may store static data or instructions required by processor 520 or other computer modules. Permanent storage may be a readable and writable storage device. Permanent storage may be a non-volatile storage device that retains stored instructions and data even when the computer is powered off. In some embodiments, the permanent storage device utilizes a mass storage device (e.g., a magnetic or optical disk, flash memory). In other embodiments, the permanent storage device may be a removable storage device (e.g., a floppy disk, optical drive). System memory may be a readable and writable storage device or a volatile readable and writable storage device, such as dynamic random access memory (DRAM). System memory may store some or all instructions and data required by the processor during operation. Furthermore, memory 510 may include any combination of computer-readable storage media, including various types of semiconductor memory chips (e.g., DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), as well as magnetic disks and / or optical disks. In some embodiments, the memory 510 may include a readable and / or writable removable storage device, such as a compact disc (CD), a read-only digital versatile disc (e.g., DVD-ROM, double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (e.g., SD card, mini SD card, Micro-SD card, etc.), a magnetic floppy disk, etc. Computer-readable storage media do not contain carrier waves and transient electronic signals transmitted wirelessly or wired.

[0098] The memory 510 stores executable codes. When the executable codes are processed by the processor 520 , the processor 520 may execute part or all of the above-mentioned methods.

[0099] In addition, the method according to the present application may also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing some or all of the steps in the above method of the present application.

[0100] Alternatively, the present application can also be implemented as a computer-readable storage medium (or non-transitory machine-readable storage medium or machine-readable storage medium) on which executable code (or computer program or computer instruction code) is stored. When the executable code (or computer program or computer instruction code) is executed by a processor of an electronic device (or server, etc.), the processor executes part or all of the steps of the above-mentioned method according to the present application.

[0101] The embodiments of the present application have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to the technology in the market, or to enable other persons skilled in the art to understand the embodiments disclosed herein.

Claims

1. An information feedback method, characterized in that: include: In response to receiving user input information, obtaining a query result from a target database based on the input information; Based on the query result, processing the input information into prompt word data; Based on the prompt word data, generating feedback information for the user through a large model; The target database includes a database that is dynamically updated according to new word determination results and / or hot word determination results.

2. The method according to claim 1, characterized in that The step of processing the input information into prompt word data based on the query result includes: generating reference information based on the query result; The input information is processed into the prompt word data using the reference information.

3. The method according to claim 2, characterized in that The query result includes a first entity word and a second entity word having a corresponding relationship; and generating reference information based on the query result includes: The first entity word and the second entity word are inserted into the preset slots of the reference information in a one-to-one correspondence to generate the reference information.

4. The method according to claim 1, wherein The step of generating feedback information for the user based on the prompt word data through a large model includes: Determining user intent based on the elements of the input information; Based on the prompt word data and the user intention, feedback information for the user is generated through a large model.

5. The method according to claim 1, wherein The input information is the input information of this round; Obtaining query results from a target database based on the input information includes: Get historical conversation information; The query result is obtained from the target database based on the current round input information and the historical conversation information.

6. The method according to claim 1, characterized in that The method further comprises: In response to receiving the timeliness data change task processing list, determining a new word determination result and / or a hot word determination result according to the timeliness data change task processing list; The entity words in the target database are updated according to the new word determination result and / or the hot word determination result.

7. The method according to claim 6, characterized in that The method further comprises: Obtain target data through real-time data synchronization processing; Based on the target data, using a hot word calculation algorithm and / or a new word calculation algorithm to obtain the new word determination result and / or the hot word determination result; and The timeliness data change task processing list is generated according to the new word determination result and / or the hot word determination result.

8. An information feedback device, characterized in that: include: an obtaining module, configured to obtain a query result from a target database based on the input information received from the user; A processing module, configured to process the input information into prompt word data based on the query result; A generation module, configured to generate feedback information for the user through a large model based on the prompt word data; The target database includes a database that is dynamically updated according to new word determination results and / or hot word determination results.

9. An electronic device, characterized in that: include: processor; as well as A memory having executable codes stored thereon, which, when executed by the processor, causes the processor to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having executable codes stored thereon, wherein when the executable codes are executed by a processor of an electronic device, the processor is caused to execute the method according to any one of claims 1 to 7.

Images