Method, device, medium and product for determining recommendation information based on task description information

Abstract

This invention discloses a method, device, medium, and product for determining recommendation information based on task description information, relating to the fields of big data and artificial intelligence. The method includes: determining the user's target task description information and performing feature information extraction processing on the target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of at least two task dimensions; determining a demand instruction template based on the task intent and at least two task dimensions, and determining a target demand instruction based on the demand instruction template and the dimensional data of at least two task dimensions; processing the target demand instruction, task description level, and task intent using an information processing model to obtain target recommendation information, and controlling an information interaction device to display the target recommendation information. This invention can quickly generate high-quality demand instructions that conform to the user's task description information and obtain recommendation information that matches the user's task description information.

Description

Technical Field

[0001] This invention relates to the fields of big data and artificial intelligence technology, and in particular to a method, device, medium and product for determining recommendation information based on task description information. Background Technology

[0002] The prompt instruction is a requirement instruction used to guide an artificial intelligence model to generate recommendation information. The artificial intelligence model can filter and organize database information according to the prompt instruction to obtain recommendation information. Therefore, the information quality of the recommendation information is closely related to the quality of the prompt instruction.

[0003] Currently, prompt instructions need to be written by users based on their own task description information. The quality of prompt instructions is greatly affected by the user's instruction writing ability and language expression. For the same task description information, different users write different prompt instructions. The uniformity, accuracy and generation speed of prompt instructions are limited, and the information quality and generation speed of recommended information are also low.

[0004] Therefore, how to quickly generate high-quality prompt instructions, improve the quality and generation speed of recommendation information, make the recommendation information more consistent with the user's task description information, and improve the user's information acquisition experience is an urgent problem to be solved. Summary of the Invention

[0005] This invention provides a method, device, medium, and product for determining recommendation information based on task description information. It can quickly generate high-quality prompt instructions that match the user's task description information, obtain recommendation information that fits the user's task description information, improve the quality and generation speed of recommendation information, and enhance the user's information acquisition experience.

[0006] According to one aspect of the present invention, a method for determining recommendation information based on task description information is provided, the method comprising: Determine the user's target task description information, and perform feature information extraction processing on the target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of at least two task dimensions; Based on the task intent and at least two task dimensions, determine the requirement instruction template, and based on the requirement instruction template and the dimensional data of at least two task dimensions, determine the target requirement instruction. Using an information processing model, the target requirement instructions, task description level, and task intent are processed to obtain target recommendation information, and the information interaction device is controlled to display the target recommendation information.

[0007] According to another aspect of the present invention, an apparatus for determining recommendation information based on task description information is provided. This apparatus is used to implement the method for determining recommendation information based on task description information in any embodiment of the present invention. The apparatus includes: The task information determination module is used to determine the user's target task description information and perform feature information extraction processing on the target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of at least two task dimensions. The requirement instruction determination module is used to determine the requirement instruction template based on the task intent and at least two task dimensions, and to determine the target requirement instruction based on the requirement instruction template and the dimension data of at least two task dimensions. The recommendation information determination module is used to process the target requirement instructions, task description level and task intent using an information processing model to obtain target recommendation information, and control the information interaction device to display the target recommendation information.

[0008] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: At least one processor; and a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by at least one processor, which enables the at least one processor to perform the method for determining recommendation information based on task description information in any embodiment of the present invention.

[0009] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the method for determining recommendation information based on task description information in any embodiment of the present invention.

[0010] According to another aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements a method for determining recommendation information based on task description information according to any embodiment of the present invention.

[0011] The method for determining recommendation information based on task description information according to the present invention includes: determining the user's target task description information, and performing feature information extraction processing on the target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of at least two task dimensions; determining a demand instruction template based on the task intent and at least two task dimensions, and determining a target demand instruction based on the demand instruction template and the dimensional data of at least two task dimensions; processing the target demand instruction, task description level, and task intent using an information processing model to obtain target recommendation information, and controlling an information interaction device to display the target recommendation information. The technical solution of the present invention parses the user's target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of at least two task dimensions; determines a demand instruction template based on the task intent and at least two task dimensions; and fills the demand instruction template using the dimensional data of at least two task dimensions to quickly generate high-quality demand instructions (e.g., prompt instructions) that conform to the user's task description information. This allows the information processing model to organize recommendation information that matches the user's task description information based on the demand instruction, task description level, and task intent, effectively improving the generation quality and speed of recommendation information and enhancing the user's information acquisition experience. This addresses the issue that the quality of prompt instructions is greatly affected by the user's instruction writing ability and language expression. Different users write different prompt instructions for the same task description information, resulting in limited uniformity, accuracy, and generation speed of prompt instructions, and consequently, low information quality and low generation speed of recommended information.

[0012] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in this invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart illustrating a method for determining recommendation information based on task description information provided by the present invention; Figure 2 This is a flowchart illustrating another method for determining recommendation information based on task description information provided by the present invention; Figure 3This is a schematic diagram of the structure of a device for determining recommendation information based on task description information provided by the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided by the present invention. Detailed Implementation

[0015] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort should fall within the scope of protection of the present invention.

[0016] It should be noted that the terms "first," "second," "initial," "intermediate," "candidate," "alternate," "target," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] The acquisition, storage, use, and processing of data in the technical solution of this invention all comply with relevant national laws and regulations. Specifically, the user information collected in this invention is information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with relevant national and regional laws, regulations, and standards, and necessary confidentiality measures are taken. This does not violate public order and good morals, and corresponding operation entry points are provided for users to choose to authorize or reject automated decision-making results; if the user chooses to reject, the process proceeds to the expert decision-making process. It should be noted that certain software, components, models, and other existing industry solutions may be mentioned in the embodiments of this application. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used the relevant content of such solutions.

[0018] High-quality prompts are one of the main ways to guide large language models to generate effective recommendation information. However, writing high-quality prompts manually requires writers with rich professional knowledge and experience. Those lacking relevant professional backgrounds will find it difficult to write prompts that can guide the model to output accurate and effective results. Specifically, different domains have different requirements for prompts. For example, in the medical field, prompts need to accurately cover medical terminology, symptom descriptions, and diagnostic procedures; in the financial field, prompts need to include market data, investment strategies, and risk assessments. Secondly, different writers vary in their language expression skills, depth of understanding of the task, and grasp of the model's characteristics. For the same text classification task, some writers may simply describe the task, such as "classify the text," while experienced writers will detail the classification criteria, category ranges, and expected output format, such as "classify the text into four categories based on news type: politics, economics, sports, and entertainment. The output format is: [text, text category]." Therefore, different writers have different levels of prompt writing skills, resulting in inconsistent prompt quality and making it difficult to guarantee the quality of generated recommendation information.

[0019] Current prompt writing methods suffer from the following problems: 1) They rely heavily on the writer's ability, resulting in low prompt generation efficiency. In practical applications, especially in scenarios involving large-scale data processing or high real-time requirements, inefficient prompt generation methods struggle to meet the demands of rapidly evolving applications. For example, in intelligent customer service systems, it is necessary to generate appropriate prompts and obtain answer information in real time based on a large number of different user questions. Manual writing cannot complete such a heavy task in a short time, easily leading to response delays and impacting user experience. Furthermore, when prompt optimization is needed, the process of repeated manual modification and testing is lengthy, which is incompatible with the rapid iteration and development of applications. 2) They rely heavily on the user's ability to express their needs and the writer's understanding and reproducibility, resulting in low prompt accuracy. If the user's expressed needs are inaccurate, or the writer's understanding or reproducibility is limited—for example, if the writer fails to accurately capture the user's true intent, or if the user's expression of needs is vague or incomplete—the generated prompt may deviate significantly from the user's actual needs, failing to guide the model to generate results that meet the user's expectations. Specifically, if a user wants to generate a "feasibility analysis report on the layout planning of electric vehicle charging stations in urban transportation systems," but the user's description is simply "write a traffic report," the writer will not understand the user's specific focus on the layout planning of electric vehicle charging stations. The generated prompt will not include the requirement for the layout planning of electric vehicle charging stations, and the feasibility analysis report output by the model will ultimately deviate from the user's actual needs.

[0020] Figure 1 This is a flowchart illustrating a method for determining recommendation information based on task description information provided by the present invention. This embodiment can deeply understand and parse the user's task description information, and generate information demand instructions in a standardized, compliant, fast, high-quality, and highly consistent manner based on the user's task description information, so as to quickly and accurately determine recommendation information and improve the user's information acquisition experience. This method can be executed by the device provided by the present invention for determining recommendation information based on task description information. This device can be implemented in hardware and / or software. In a specific embodiment, the device can be integrated into an electronic device. The following embodiments will illustrate this using the integration of the device into an electronic device as an example. Figure 1 The method specifically includes the following steps: S101. Determine the user's target task description information, and perform feature information extraction processing on the target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of at least two task dimensions.

[0021] The user's target task description information can be understood as the specific goal or information the user hopes to achieve through the prompt command, such as querying medical data, generating travel guide documents, or generating images. This target task description information is a text that accurately, comprehensively, and multidimensionally expresses the user's information usage needs. Depending on the comprehensiveness and accuracy of the user's feedback, the target task description information can be determined directly based on the information usage needs, or it can be determined after at least one round of questioning with the user based on those needs. Secondly, this invention does not limit users to providing feedback on their information usage needs in text form; it can also be in the form of audio, etc. For example, when it is inconvenient for users to input text, voice interaction devices can be used to recognize the user's audio data. This setup can effectively expand the applicable scenarios of the recommendation information, improve the user experience, and enhance the practicality of the information recommendation method. If the information usage needs are provided in a non-text form such as audio data, the audio data will first be parsed and organized, then converted into text, and then analyzed to finally obtain the user's target task description information. Feature extraction processing of target task description information can be understood as extracting task feature fields from the target task description information to analyze the core content of the user's task description information. Feature fields can be fields representing the user's information usage needs, obtained by parsing the grammatical structure, lexical meaning, semantic relationships, and user sentiment tendencies of the target task description information using semantic analysis and sentiment analysis techniques. These fields can be used to determine information such as the user's task intent, the degree of task description, at least two task dimensions, and dimensional data for at least two task dimensions. Specifically, when a user inputs "I want a positive and inspiring speech," semantic analysis can reveal that the user's required emotional tone is "positive" and "inspiring," and the user's need is "a speech." Therefore, the extracted feature fields are positive, inspiring, and a speech.

[0022] Specifically, a user's task intent can be understood as their direct information usage needs, such as a speech draft, item prices, or generating cartoons. The level of user task description can be understood as the difficulty, professionalism, or applicability of the information the user needs, such as simple cartoons, complex cartoons, popular science speeches, research speeches, fourth-grade math tests, or high school physics tests. A user's task dimension can be understood as the descriptive angle of their information usage needs. The dimensional data of the task dimension can be understood as the specific descriptive content of the corresponding descriptive angle. For example, if a user's target task description is "generate five paintings suitable for popular science, with an abstract style and a human figure theme," the user's task intent is five paintings, the task description level is popular science, the task dimensions include style and theme, the style dimension's dimensional data is abstract, and the theme dimension's dimensional data is human figures. The advantage of this setup is that it comprehensively, from multiple angles, and accurately analyzes the user's task description information to generate precise, high-quality demand instructions, thus improving the accuracy of recommended information.

[0023] In one specific implementation, determining the user's target task description information includes the following four steps: A1, A2, A3, and A4.

[0024] A1. Obtain the user's initial task description information.

[0025] The user's initial task description information can be understood as the text information corresponding to the task description text or audio message sent by the user for the first time. For example, when the user's first input of the task description text or audio message is "I want to generate a painting", "I want to generate a painting" is considered as the initial task description information.

[0026] A2. Determine whether the initial task description information meets the requirements for generating the instruction.

[0027] The requirement instruction generation condition can be understood as the basis for whether the initial task description information of the beam can generate a prompt instruction that accurately expresses the user's information usage needs. A prompt instruction that accurately expresses the user's information usage needs is more likely to find recommended content that satisfies the user, while the more vague the information usage needs, the lower the likelihood of finding recommended content that satisfies the user. The purpose of this invention is to only process the task description information that can generate a prompt instruction that accurately expresses the user's information usage needs, thereby saving processor data processing volume and data processing resources while ensuring the quality of prompt generation and the effectiveness of information recommendation.

[0028] Specifically, the conditions for generating a demand instruction can be that the task description information at least covers the main characteristics of the information usage requirement and a limiting description of those main characteristics. For example, assuming the main characteristics of a painting include theme, audience, and type, then the requirement to generate the painting must at least include descriptions of these three types of information. For instance, when the initial task description information is "I want to generate a painting," it is considered that the user's task description information may be ambiguous, making it impossible to determine what type of painting the user needs, what the theme of the painting is, or what the target audience of the painting is. Based on this information, the probability of finding a painting that satisfies the user is low, and the initial task description information is considered not to meet the conditions for generating a demand instruction. When the initial task description information is "I want to generate a painting suitable for popular science, with an abstract style and a portrait theme," it is considered that the initial task description information is relatively clear, and the probability of finding a painting that satisfies the user is high, and the initial task description information is considered to meet the conditions for generating a demand instruction.

[0029] A3. If the initial task description information meets the conditions for generating the required instruction, then the initial task description information is determined to be the target task description information.

[0030] A4. If the initial task description information does not meet the conditions for generating the required instruction, then the initial task description information is semantically extracted to obtain the user's task preference. Based on the task preference, a task description information guidance instruction is generated, and the information interaction device is controlled to display the task description information guidance instruction. Supplementary task description information fed back by the user based on the task description information guidance instruction is obtained. The task description information set of the supplementary task description information and the initial task description information is used to update the initial task description information, and the process returns to the step of determining whether the initial task description information meets the conditions for generating the required instruction.

[0031] Among them, the user's task preference can be understood as the user's information usage needs. It can be used to determine the type of task description information of the user and the main characteristics of that type of task description information, and then determine the content that is missing in the initial task description information, so as to guide the user to supplement the information and quickly obtain accurate and comprehensive task description information. For example, if a user's initial task description is "I want to generate a painting," then the user's task description is determined to be a painting. The main characteristics of a painting are theme, audience, and type. However, none of these are present in the initial task description. Therefore, the user is guided to supplement the theme, audience, and type information. The task description guidance instructions are designed to guide the user to supplement the theme, audience, and type information. For example, "What style do you want this painting to be? Realistic, abstract, or cartoon style?", "What theme do you want this painting to be? Landscape, portrait, or animal?", "Who do you want the target audience for this painting to be? People who work in the art world or other groups?" Supplementing the task description information can be understood as the user providing secondary feedback on the task description information after viewing the task description guidance instructions. After receiving the supplemented task description information, the task description information set of the supplemented task description information and the initial task description information is determined as the new initial task description information. Further analysis is then conducted to determine whether the new task description information meets the conditions for generating the required instruction.

[0032] It's worth noting that if the user's supplementary task description details the painting's theme, target audience, and type, the updated task description will meet the requirements for generating the demand instruction. However, if the user's supplementary task description lacks these three aspects, or if the description is incomplete or missing, the updated task description will not meet the requirements for generating the demand instruction. In this case, further questioning and updating of the initial task description are necessary until the updated initial task description covers all three aspects. This design aims to guide users to improve their task description through at least one round of questioning when the initial demand information is inadequate, thereby generating accurate demand instructions and maximizing the probability that recommended information matches the user's task description, ultimately enhancing the user's information retrieval experience.

[0033] S102. Based on the task intent and at least two task dimensions, determine the requirement instruction template, and based on the requirement instruction template and the dimension data of at least two task dimensions, determine the target requirement instruction.

[0034] The requirement instruction template can be understood as a general framework of requirement instructions determined based on task intent and at least two task dimensions. The target requirement instruction can be understood as a requirement instruction that fits the user's information usage needs after being filled with the user's personalized task description data (i.e., dimensional data of at least two task dimensions). Based on the target requirement instruction, content that meets the user's information usage needs can be obtained with high precision.

[0035] On the one hand, based on the task intent and at least two task dimensions, a requirement instruction template is determined, including: constructing a task description information vector based on the task intent and at least two task dimensions; and using the task description information vector to filter from a pre-built requirement instruction template library to obtain the requirement instruction template.

[0036] A pre-built requirement instruction template library can be understood as a collection of general requirement instruction templates obtained by summarizing, classifying, and refining various types of requirement instructions. It covers a variety of task description information instructions across multiple fields, including natural language processing tasks such as text generation, question-answering systems, machine translation, and image generation. It also includes templates designed to improve the efficiency of identifying target requirement instructions through template filtering. The personalized task description data for instructions in the requirement instruction template library is initially empty; it needs to be filled in using user information to obtain the user's personalized target requirement instructions. The task description information vector can be understood as a mathematical representation of the task description information obtained by integrating and mathematically transforming the task intent and at least two task dimensions. This design facilitates the computer's understanding and processing of the user's task description information, enabling it to quickly select requirement instruction templates from the pre-built requirement instruction template library that match the task description information vector.

[0037] Specifically, the requirement instruction template includes at least two candidate instruction templates. These candidate templates can be any instruction framework in a pre-built requirement instruction template library that matches the task description information vector, or they can be the instruction frameworks in the pre-built requirement instruction template library that rank in the top X (X is an integer not less than 2) positions according to their matching degree with the task description information vector. Each candidate instruction template can realize the user's target task description information. The purpose of setting multiple candidate instruction templates is to filter out templates that match the target task description information, so as to determine multiple backup requirement instructions based on each candidate instruction template and the dimensional data of at least two task dimensions. Then, based on the detection results of each backup requirement instruction, the optimal requirement instruction is determined as the target requirement instruction, thus improving the instructions of the target requirement instruction and the recommended information. Alternatively, based on the detection results of each backup requirement instruction, the backup requirement instructions can be continuously optimized, and the backup requirement instructions with better adaptability, universality, compliance, and expressiveness can be selected as the target requirement instruction, ensuring that the recommended information meets the user's information usage needs while improving the instructions of the recommended information as much as possible.

[0038] On the other hand, based on the requirement instruction template and dimensional data of at least two task dimensions, the target requirement instruction is determined, including the following three steps: B1, B2, and B3.

[0039] B1. Based on the principle of dimension correspondence and the dimensional data of at least two task dimensions, fill in at least two candidate instruction templates to obtain at least two backup requirement instructions.

[0040] The backup requirement instructions and candidate instruction templates are in a one-to-one correspondence. Each candidate instruction template includes multiple sets of sub-instructions, each corresponding to a task dimension. When filling in the instruction framework, the corresponding dimension data must be filled into the blank spaces of the sub-instructions. That is, according to the dimension correspondence principle, dimension data of dimension A is adaptively filled into the blank spaces of the sub-instructions corresponding to dimension A, and dimension data of dimension B is adaptively filled into the blank spaces of the sub-instructions corresponding to dimension B. This setup aims to ensure the accuracy of the backup requirement instructions.

[0041] B2. Perform instruction performance verification processing on at least two backup requirement instructions to obtain the instruction performance verification results for at least two backup requirement instructions.

[0042] The purpose of instruction performance verification is to verify the compliance and functional implementation of backup requirement instructions, so as to select the best-performing backup requirement instructions or improve the performance of backup requirement instructions by adjusting the parameters of each backup requirement instruction. The instruction performance verification result of backup requirement instructions can be understood as the performance test result of backup requirement instructions. For example, if it is necessary to perform compliance testing and instruction content evaluation on backup requirement instructions, then the instruction performance verification result of backup requirement instructions includes compliance testing results and content evaluation results.

[0043] B3. When it is determined, based on the instruction performance verification results of at least two backup requirement instructions, that there is an instruction among the at least two backup requirement instructions that meets the conditions for generating recommendation information, the instruction that meets the conditions for generating recommendation information is determined as the target requirement instruction; when it is determined, based on the instruction performance verification results of at least two backup requirement instructions, that there is no instruction among the at least two backup requirement instructions that meets the conditions for generating recommendation information, the at least two backup requirement instructions are adjusted according to the instruction performance verification results of at least two backup requirement instructions, and the process is returned to perform instruction performance verification processing on at least two backup requirement instructions to obtain the instruction performance verification results of at least two backup requirement instructions.

[0044] The recommendation information generation conditions can be understood as the criteria for determining whether backup demand instructions are qualified. For example, the instruction performance verification results of backup demand instructions include whether the compliance test results are compliant and whether the content evaluation results are accurate and comprehensive. If there are instructions among the backup demand instructions that meet the recommendation information generation conditions, the instruction can be directly identified as the target demand instruction, thus improving the generation rate of recommendation information. If there are no instructions among the backup demand instructions that meet the recommendation information generation conditions, at least two backup demand instructions can be optimized based on the instruction performance verification results of at least two backup demand instructions. Then, the optimized backup demand instructions can be further subjected to instruction performance verification processing until an instruction that meets the recommendation information generation conditions is obtained, thereby ensuring the accuracy of the generated recommendation information.

[0045] In one specific implementation, a genetic algorithm can be used to simulate the inheritance, mutation, and selection mechanisms in biological evolution to optimize the prompt. The initially generated multiple prompts (i.e., at least two initially determined backup demand instructions) are regarded as individuals in a biological population. New prompt individuals are generated through crossover operations (i.e., combining parts of different prompts) and mutation operations (i.e., randomly changing parts of the prompts). Then, based on evaluation indicators, prompt individuals with high fitness (i.e., good generation effect) are continuously selected, while prompt individuals with low fitness are eliminated. After multiple rounds of iteration, a better prompt is obtained and used as the target demand instruction.

[0046] S103. Using an information processing model, process the target requirement instructions, task description level, and task intent to obtain target recommendation information, and control the information interaction device to display the target recommendation information.

[0047] The information processing model can be understood as an algorithm that can adaptively select and summarize information based on demand instructions. The target recommendation information can be understood as the information obtained by the information processing model after processing the target demand instructions, the degree of task description, and the task intent. The information interaction device can be understood as a device that displays information and obtains the user's task description information, so that the user can input and provide feedback on the task description information and visually display the target recommendation information, thereby improving the user's information acquisition experience.

[0048] Using an information processing model, the target requirement instructions, the degree of task description, and the task intent are processed. For example, a knowledge base for recommended information can be determined based on the degree of task description and the task intent. Using the information processing model, the information in the knowledge base of recommended information is filtered, processed, and integrated based on the target requirement instructions to obtain the summary information. This summary information includes, but is not limited to, text, images, tables, and formulas.

[0049] In one specific implementation, an information processing model is used to process the target requirement instructions, the degree of task description, and the task intent to obtain target recommendation information, including the following three steps C1, C2, and C3.

[0050] C1. Determine the initial recommendation information based on the degree of task description and the information task intent.

[0051] The initial recommendation information can be understood as the source dataset for recommendation information. The information task intent is used to determine the type database of recommendation information, and the task description level is used to determine the supplementary database of recommendation information. The initial recommendation information is a set of information from the supplementary information database and the type database. It is worth noting that information deduplication is performed when determining the information set to reduce the amount of information while ensuring comprehensiveness and improving information processing efficiency.

[0052] For example, if the information task intent is medical, then the type database is a medical database, containing information such as concepts, entities, relationships, diseases, symptoms, treatment methods, and drugs in the medical field. If the task description level is popular science, then the supplementary information database is empty or a simple medical knowledge base. If the task description level is research, then the supplementary information database includes, but is not limited to, a database of academic papers and industry reports in the medical field, so as to parse professional terminology, symptom manifestations, diagnostic criteria, and other knowledge of diseases from the supplementary information database, thereby improving the information acquisition satisfaction of users with in-depth search needs.

[0053] It is worth noting that the level of task description can be used to filter information sources, as well as to generate demand instructions. That is, the level of task description can be directly integrated into demand instructions to obtain instructions that better express the user's information usage needs. By improving the quality of demand instructions, the quality of recommended information can be improved.

[0054] C2. Based on the initial recommendation information, target demand instructions, and information processing model, determine the intermediate recommendation information.

[0055] Intermediate recommendation information can be understood as the summary information obtained by using an information processing model to filter, process, and integrate the initial recommendation information based on the target demand instructions.

[0056] C3. Evaluate the intermediate recommendation information to obtain the evaluation results. When it is determined that the intermediate recommendation information meets the conditions for displaying recommendation information based on the evaluation results, the intermediate recommendation information is determined as the target recommendation information.

[0057] The evaluation of intermediate recommendation information can be understood as diagnosing and evaluating it. This includes similarity assessment, accuracy assessment, logical consistency assessment, and richness assessment. Similarity assessment evaluates the degree of similarity between the intermediate recommendation information and the user task description information; accuracy assessment evaluates whether the intermediate recommendation information covers the necessary content corresponding to the information usage needs; logical consistency assessment determines whether the logic of the intermediate recommendation information is self-consistent; and richness assessment evaluates the multi-dimensional description of the intermediate recommendation information. The evaluation results of the intermediate recommendation information can be understood as the conclusions of each diagnostic assessment, including similarity assessment results, accuracy assessment results, logical consistency assessment results, and richness assessment results. The display criteria for recommended information can be understood as the basis for determining whether recommended information meets the user task description information and the information rationality rules. The information rationality rules are used to determine whether the information is compliant, logically consistent, and contains multiple dimensions. If intermediate recommended information meets the display criteria—for example, if the similarity between the intermediate recommended information and the target task description information is higher than 80% (the specific value is not limited), the intermediate recommended information includes the necessary content corresponding to the information usage requirements, there are no logical problems, and it includes descriptions of at least two dimensions—then it is proven that the intermediate recommended information is rational information that conforms to the user task description information, and it will be determined as the target recommended information. Conversely, if the intermediate recommended information does not meet the display criteria, the target requirement instruction is readjusted based on the evaluation results of the intermediate recommended information, and the target recommended information continues to be generated to obtain rational information that conforms to the user task description information. The advantage of this setup is that it allows for rapid verification of intermediate recommended information based on the evaluation results, improving the efficiency of recommending information determination.

[0058] Specifically, the intermediate recommendation information is evaluated to obtain the evaluation results, including: determining the similarity evaluation result based on the cosine similarity calculation rule (i.e., the cosine similarity algorithm), the intermediate recommendation information, and the target task description information, which is equivalent to using the cosine similarity algorithm to calculate the matching degree between the intermediate recommendation information and the target task description information; determining at least one recommendation information verification text based on the task intent and the degree of task description, and determining the accuracy evaluation result based on the intermediate recommendation information and at least one recommendation information verification text; performing compliance coherence detection on the intermediate recommendation information, and determining the logicality evaluation result based on the detection results; determining the dimensional data of the intermediate recommendation information, and determining the richness evaluation result based on the dimensional data of the intermediate recommendation information.

[0059] Among them, the recommendation information verification text can be understood as the necessary information corresponding to the information usage requirements. This information can be determined according to the task intent and belongs to general verification information. It is used to detect whether the intermediate recommendation information has complete necessary feature information. The accuracy evaluation result can be understood as the conclusion indicating whether the intermediate recommendation information has complete necessary feature information. The compliance coherence test is used to determine whether there are logical conflicts and contradictions in the intermediate recommendation information. The accuracy evaluation result can be understood as the compliance coherence test conclusion of the intermediate recommendation information. The richness evaluation result can be understood as the dimensional description conclusion of the intermediate recommendation information. The more dimensional descriptions of the intermediate recommendation information, the better the richness evaluation result.

[0060] It is worth noting that after controlling the information interaction device to display the target recommendation information, the method of the present invention further includes: receiving recommendation information evaluation information triggered by the user based on the information interaction device; determining whether the recommendation information evaluation information contains a recommendation information adjustment instruction; and if the recommendation information evaluation information contains a recommendation information adjustment instruction, adjusting the target demand instruction based on the recommendation information adjustment instruction.

[0061] Recommendation information evaluation information can be understood as user feedback after viewing recommended information. Users may be satisfied or dissatisfied with the recommended information. If a user is dissatisfied with the target recommended information, this invention supports regenerating the demand instruction based on the reason for the user's dissatisfaction to adjust the recommended information. This repeated feedback adjustment process yields recommended information that satisfies the user. Generally, when a user is satisfied with the target recommended information, the recommendation information evaluation information does not include adjustment instructions. When a user is dissatisfied, the recommendation information evaluation information includes adjustment instructions. These adjustment instructions can be understood as the reason for the user's dissatisfaction with the target recommended information and their adjustment needs, such as "The recommended information is too technical; please make it easier to understand," or "The recommended information is too simplistic; please make it more professional." This allows for the adjustment of the target demand instruction based on the adjustment instructions. The advantage of this setup is that it records the user's information acquisition preferences, enabling subsequent processing of the user's information acquisition needs to adapt and adjust the demand instructions according to these preferences, resulting in more satisfactory recommended information.

[0062] In one specific implementation, the information recommendation process may include: 1) The user inputs task description information via text input. For example, if the user inputs "I want to create a travel guide," and the system recognizes this as a simple and vague travel guide creation request, it will begin interacting with the user through an intelligent follow-up questioning strategy to further inquire about the user's travel guide creation requirements. For example, it may ask the user, "Which travel destination do you want to visit? When do you plan to travel? What is your approximate budget?" 2) If the user replies, "I want to travel to location A in mid-July, with a budget of approximately X yuan," it can continue to ask the user, "Which attractions in location A are you most interested in? Are they natural scenery, such as attraction 1 and attraction 2, or cultural attractions, such as attraction 3 and attraction 4?" 3) If the user's answer is more inclined towards natural scenery attractions, semantic analysis and sentiment analysis techniques will be used to analyze the task description information. For example, the core of understanding the user's needs is to create a travel guide focusing on natural scenery attractions in location A, for a trip in July with a budget of X yuan. These key elements will be extracted to construct a demand vector. Then, a travel guide template will be selected from the prompt template library based on the constructed demand vector. This template includes key elements such as attraction introduction, itinerary, and precautions. 4) Key information from user needs, such as the travel destination (location A), travel time (July), budget (X yuan), and preference for natural scenery, is populated into a template to generate an initial prompt. This initial prompt is then optimized using reinforcement learning and genetic algorithms. For example, during reinforcement learning, the generation strategy is adjusted based on feedback from previously generated travel guides (e.g., user evaluations of the itinerary's rationality and the richness of the attraction descriptions). In terms of genetic algorithms, the initially generated prompts are treated as individuals in a population, and new prompts are generated through crossover and mutation operations. After multiple iterations, the prompt with the best performance is selected as the target requirement instruction. Furthermore, knowledge bases such as professional literature databases and tourism knowledge graphs for location A can be retrieved to incorporate information such as July weather characteristics, detailed information about attractions, and precautions for the travel season into the final prompt, thus improving the quality of the recommended information. 5) The large model generates travel guides based on the generated prompts and evaluates them from multiple perspectives, including content completeness, relevance, practicality, and logic. For example, the content completeness dimension checks whether the guide covers the user's needs for information such as attraction introductions, itinerary arrangements, and precautions; the relevance dimension judges whether the attraction introductions and itinerary arrangements are closely related to the trip to location A in July and are within the budget of X yuan; the practicality dimension assesses whether the itinerary arrangements are reasonable (e.g., whether the transportation connections are smooth, whether the accommodation choices are within the budget, and whether the accommodations are convenient for sightseeing); and the logic dimension checks whether the structure of the guide is clear.6) Show the generated travel guide to users and determine if adjustments are needed based on their feedback. For example, if a user suggests, "I'd like a more relaxed itinerary, less rushed, and with more recommendations for local delicacies," the prompt will be revised to include rest periods in the itinerary section and local food recommendations after the notes. The travel guide will then be regenerated, and user feedback will be collected until the user is satisfied. This process can also record user feedback and learn their preferences for relaxed itineraries and food recommendations to automatically optimize the prompt when generating similar content in the future.

[0063] The technical solution of the above embodiments parses the user's target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of at least two task dimensions. Based on the task intent and at least two task dimensions, a requirement instruction template is determined, and the dimensional data of at least two task dimensions is used to populate the requirement instruction template, quickly generating high-quality prompt instructions that match the user's task description information. This allows the information processing model to organize recommended information that fits the user's task description information based on the requirement instruction, task description level, and task intent, effectively improving the quality and speed of recommended information generation and enhancing the user's information acquisition experience. Secondly, multimodal interaction and intelligent question answering can be used to deeply mine the user's task description information, improving the accuracy of understanding the user's task description information and avoiding the impact of inaccurate user demand expression or inadequate understanding on the quality of requirement instructions. Furthermore, the requirement instructions are iterated and optimized during generation to further improve their accuracy and quality, ensuring that high-quality recommended information that better matches the user's needs is obtained. This addresses the issue that the quality of prompt instructions is greatly affected by the user's instruction writing ability and language expression. Different users write different prompt instructions for the same task description information, resulting in limited uniformity, accuracy, and generation speed of prompt instructions, and consequently, low information quality and low generation speed of recommended information.

[0064] Figure 2 This is a flowchart illustrating another method for determining recommendation information based on task description information provided by the present invention. Based on the above embodiments, this embodiment provides a preferred method for determining demand instructions and recommendation information with more complete process details. Specifically, as shown... Figure 2 As shown, the method includes: S201. Obtain the user's initial task description information.

[0065] S202. Determine whether the initial task description information meets the requirements for generating the instruction.

[0066] Specifically, if the initial task description information meets the conditions for generating the required instruction, then S203 is executed; otherwise, if the initial task description information does not meet the conditions for generating the required instruction, the initial task description information is semantically extracted to obtain the user's task preference. Based on the task preference, a task description information guidance instruction is generated, and the information interaction device is controlled to display the task description information guidance instruction. Supplementary task description information fed back by the user based on the task description information guidance instruction is obtained. The initial task description information is updated using the task description information set of the supplementary task description information and the initial task description information, and the process returns to the step of determining whether the initial task description information meets the conditions for generating the required instruction, that is, S202 is executed again.

[0067] S203. Determine the initial task description information as the target task description information.

[0068] S204. Perform feature information extraction processing on the target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of at least two task dimensions.

[0069] S205. Construct a task description information vector based on the task intent and at least two task dimensions.

[0070] S206. Filter the pre-built requirement instruction template library using the task description information vector to obtain the requirement instruction template.

[0071] S207. Based on the principle of dimension correspondence and the dimension data of at least two task dimensions, fill in at least two candidate instruction templates to obtain at least two backup requirement instructions.

[0072] Among them, the backup demand instructions and the candidate instruction templates correspond one-to-one.

[0073] S208. Perform instruction performance verification processing on at least two backup requirement instructions to obtain instruction performance verification results for at least two backup requirement instructions.

[0074] S209. When it is determined, based on the instruction performance verification results of at least two backup requirement instructions, that there is an instruction among the at least two backup requirement instructions that meets the recommendation information generation conditions, the instruction that meets the recommendation information generation conditions is determined as the target requirement instruction.

[0075] Specifically, when it is determined, based on the instruction performance verification results of at least two backup requirement instructions, that there is no instruction among the at least two backup requirement instructions that meets the conditions for generating recommendation information, the at least two backup requirement instructions are adjusted according to the instruction performance verification results of the at least two backup requirement instructions, and the process returns to the step of performing instruction performance verification processing on the at least two backup requirement instructions to obtain the instruction performance verification results of the at least two backup requirement instructions, that is, returning to execute S208.

[0076] S210. Determine the initial recommendation information based on the degree of task description and the information task intent.

[0077] S211. Based on the initial recommendation information, target demand instructions, and information processing model, determine the intermediate recommendation information.

[0078] S212. Evaluate the intermediate recommendation information to obtain the evaluation result of the intermediate recommendation information, and determine the intermediate recommendation information as the target recommendation information when it is determined that the intermediate recommendation information meets the conditions for displaying recommendation information based on the evaluation result of the intermediate recommendation information.

[0079] Specifically, after determining the target recommendation information, the system will also control the information interaction device to display the target recommendation information so that users can view and understand the recommended information and adjust the recommended information by changing instructions.

[0080] Figure 3 This is a schematic diagram of a device for determining recommendation information based on task description information provided by the present invention. Figure 3 As shown, the device includes: a task information determination module 301, a demand instruction determination module 302, and a recommendation information determination module 303.

[0081] The task information determination module 301 is used to determine the user's target task description information and perform feature information extraction processing on the target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of at least two task dimensions.

[0082] The requirement instruction determination module 302 is used to determine the requirement instruction template based on the task intent and at least two task dimensions, and to determine the target requirement instruction based on the requirement instruction template and the dimension data of at least two task dimensions.

[0083] The recommendation information determination module 303 is used to process the target requirement instruction, task description level and task intent using an information processing model to obtain target recommendation information, and control the information interaction device to display the target recommendation information.

[0084] Optionally, the task information determination module 301 is specifically used for: obtaining the user's initial task description information; determining whether the initial task description information meets the conditions for generating a required instruction; if the initial task description information meets the conditions for generating a required instruction, then determining the initial task description information as the target task description information; if the initial task description information does not meet the conditions for generating a required instruction, then performing semantic extraction processing on the initial task description information to obtain the user's task preference, generating a task description information guidance instruction based on the task preference, and controlling the information interaction device to display the task description information guidance instruction; obtaining supplementary task description information fed back by the user based on the task description information guidance instruction, updating the initial task description information using the task description information set of the supplementary task description information and the initial task description information, and returning to the step of determining whether the initial task description information meets the conditions for generating a required instruction.

[0085] Optionally, the requirement instruction determination module 302 is specifically used to: construct a task description information vector based on the task intent and at least two task dimensions; and use the task description information vector to filter from a pre-built requirement instruction template library to obtain a requirement instruction template.

[0086] Optionally, the requirement instruction template includes at least two candidate instruction templates. The requirement instruction determination module 302 is specifically used for: filling the at least two candidate instruction templates based on the dimension correspondence principle and the dimensional data of at least two task dimensions to obtain at least two backup requirement instructions, wherein the backup requirement instructions and the candidate instruction templates correspond one-to-one; performing instruction performance verification processing on the at least two backup requirement instructions to obtain the instruction performance verification results of the at least two backup requirement instructions; when it is determined from the instruction performance verification results of the at least two backup requirement instructions that there is an instruction that meets the recommendation information generation conditions, the instruction that meets the recommendation information generation conditions is determined as the target requirement instruction; when it is determined from the instruction performance verification results of the at least two backup requirement instructions that there is no instruction that meets the recommendation information generation conditions, the at least two backup requirement instructions are adjusted according to the instruction performance verification results of the at least two backup requirement instructions, and the process is returned to the step of performing instruction performance verification processing on the at least two backup requirement instructions to obtain the instruction performance verification results of the at least two backup requirement instructions.

[0087] Optionally, the recommendation information determination module 303 is specifically used for: determining initial recommendation information based on the task description level and information task intent; determining intermediate recommendation information based on the initial recommendation information, target requirement instructions, and information processing model; evaluating the intermediate recommendation information to obtain the evaluation result of the intermediate recommendation information; and determining the intermediate recommendation information as the target recommendation information when the evaluation result of the intermediate recommendation information determines that the intermediate recommendation information meets the recommendation information display conditions.

[0088] Optionally, the evaluation results of intermediate recommendation information include similarity evaluation results, accuracy evaluation results, logicality evaluation results, and richness evaluation results; the recommendation information determination module 303 is specifically used to: determine the similarity evaluation result based on the cosine similarity calculation rule, intermediate recommendation information, and target task description information; determine at least one recommendation information verification text based on task intent and task description degree, and determine the accuracy evaluation result based on the intermediate recommendation information and at least one recommendation information verification text; perform compliance coherence detection on the intermediate recommendation information, and determine the logicality evaluation result based on the detection result; determine the dimensional data of the intermediate recommendation information, and determine the richness evaluation result based on the dimensional data of the intermediate recommendation information.

[0089] Optionally, the device for determining recommendation information based on task description information further includes an instruction adjustment module, which is used to receive recommendation information evaluation information triggered by the user based on the information interaction device after the control information interaction device displays the target recommendation information; determine whether the recommendation information evaluation information contains a recommendation information adjustment instruction; and if the recommendation information evaluation information contains a recommendation information adjustment instruction, adjust the target requirement instruction based on the recommendation information adjustment instruction.

[0090] The apparatus for determining recommendation information based on task description information provided by the present invention can execute the method for determining recommendation information based on task description information provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0091] Figure 4 This is a schematic diagram of the structure of an electronic device provided by the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0092] like Figure 4As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the read-only memory 12 or loaded from the storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, read-only memory 12, and RAM 13 are interconnected via a bus 14. An input / output interface 15 is also connected to the bus 14.

[0093] Multiple components in electronic device 10 are connected to input / output interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of monitors, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0094] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, central processing units, graphics processing units, various special-purpose artificial intelligence computing chips, various processors running machine learning model algorithms, digital signal processors, and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as methods for determining recommendation information based on task description information.

[0095] In some embodiments, the method for determining recommendation information based on task description information can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via read-only memory 12 and / or communication unit 19. When the computer program is loaded into random access memory 13 and executed by processor 11, one or more steps of the method for determining recommendation information based on task description information described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the method for determining recommendation information based on task description information by any other suitable means (e.g., by means of firmware).

[0096] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays, application-specific integrated circuits (ASICs), application-specific standard products (ASICs), systems-on-a-chip (SoCs), payload programmable logic devices, computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0097] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0098] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory / flash memory, optical fibers, portable compact disk read-only memory, optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0099] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a cathode ray tube or liquid crystal display) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (voice input and / or tactile input).

[0100] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0101] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product within the cloud computing service system to address the shortcomings of traditional physical hosts and virtual private servers, such as high management difficulty and weak business scalability.

[0102] In one specific embodiment, the present invention also includes a computer program product comprising a computer program that, when executed by a processor, implements the method for determining recommendation information based on task description information according to any embodiment of the present invention.

[0103] In the implementation of a computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​as well as conventional procedural programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including local area networks (LANs) or wide area networks (WANs), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0104] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0105] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method of determining recommendation information based on task description information, characterized by, include: Determine the user's target task description information, and perform feature information extraction processing on the target task description information to obtain the user's task intent, task description level, at least two task dimensions, and dimensional data of the at least two task dimensions; Based on the task intent and the at least two task dimensions, a requirement instruction template is determined, and based on the requirement instruction template and the dimension data of the at least two task dimensions, a target requirement instruction is determined. Using an information processing model, the target requirement instruction, the task description level, and the task intent are processed to obtain target recommendation information, and the information interaction device is controlled to display the target recommendation information.

2. The method of claim 1, wherein, The information used to determine the user's target task description includes: Obtain the user's initial task description information; Determine whether the initial task description information meets the requirements for generating the instruction; If the initial task description information meets the requirement instruction generation conditions, then the initial task description information is determined to be the target task description information; If the initial task description information does not meet the conditions for generating the required instruction, then the initial task description information is semantically extracted to obtain the user's task preference. Based on the task preference, a task description information guidance instruction is generated, and the information interaction device is controlled to display the task description information guidance instruction. Obtain supplementary task description information from the user's guidance feedback based on the task description information; update the initial task description information using the supplementary task description information and the initial task description information set; and return to the step of determining whether the initial task description information meets the requirements for generating the instruction.

3. The method of claim 1, wherein, The process of determining the requirement instruction template based on the task intent and the at least two task dimensions includes: Construct a task description information vector based on the task intent and the at least two task dimensions; The task description information vector is used to filter through a pre-built requirement instruction template library to obtain the requirement instruction template.

4. The method of claim 1, wherein, The demand instruction template includes at least two candidate instruction templates; The step of determining the target requirement instruction based on the requirement instruction template and the dimensional data of at least two task dimensions includes: Based on the principle of dimension correspondence and the dimension data of the at least two task dimensions, the at least two candidate instruction templates are filled to obtain at least two backup requirement instructions, wherein the backup requirement instructions and the candidate instruction templates correspond one-to-one. Perform instruction performance verification processing on the at least two backup requirement instructions to obtain the instruction performance verification results of the at least two backup requirement instructions; When it is determined, based on the instruction performance verification results of the at least two backup requirement instructions, that there is an instruction among the at least two backup requirement instructions that meets the recommendation information generation conditions, the instruction that meets the recommendation information generation conditions is determined as the target requirement instruction; When it is determined, based on the instruction performance verification results of the at least two backup requirement instructions, that there is no instruction among the at least two backup requirement instructions that meets the recommendation information generation conditions, the at least two backup requirement instructions are adjusted according to the instruction performance verification results of the at least two backup requirement instructions, and the process is returned to perform instruction performance verification processing on the at least two backup requirement instructions to obtain the instruction performance verification results of the at least two backup requirement instructions.

5. The method according to claim 1, characterized in that, The process of using an information processing model to process the target requirement instruction, the task description level, and the task intent to obtain target recommendation information includes: Based on the task description level and the information task intent, determine the initial recommendation information; Based on the initial recommendation information, the target demand instruction, and the information processing model, intermediate recommendation information is determined; The intermediate recommendation information is evaluated to obtain the evaluation result of the intermediate recommendation information. When it is determined that the intermediate recommendation information meets the conditions for displaying recommendation information based on the evaluation result of the intermediate recommendation information, the intermediate recommendation information is determined to be the target recommendation information.

6. The method according to claim 5, characterized in that, The evaluation results of the intermediate recommendation information include similarity evaluation results, accuracy evaluation results, logicality evaluation results, and richness evaluation results; correspondingly, the evaluation of the intermediate recommendation information to obtain the evaluation results of the intermediate recommendation information includes: The similarity evaluation result is determined based on the cosine similarity calculation rules, the intermediate recommendation information, and the target task description information; Based on the task intent and the degree of task description, at least one recommended information verification text is determined, and based on the intermediate recommended information and the at least one recommended information verification text, the accuracy evaluation result is determined; The intermediate recommendation information is subjected to compliance and coherence checks, and the logicality evaluation result is determined based on the check results. The dimensional data of the intermediate recommendation information are determined, and the richness evaluation result is determined based on the dimensional data of the intermediate recommendation information.

7. The method according to claim 1, characterized in that, After the information interaction device displays the target recommendation information, the method further includes: Accept recommendation and evaluation information triggered by the user based on the information interaction device; Determine whether the recommendation information evaluation information contains a recommendation information adjustment instruction; If the recommendation information evaluation information contains the recommendation information adjustment instruction, then the target demand instruction is adjusted based on the recommendation information adjustment instruction.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the method for determining recommendation information based on task description information as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the method for determining recommendation information based on task description information as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method for determining recommendation information based on task description information as described in any one of claims 1 to 7.

Images