Conversational exercise scheme generation method and system based on large language model, and medium

Abstract

The invention discloses a dialogue type motion scheme generation method and system based on a large language model and a medium, and relates to the field of data processing.The method comprises the steps that text vectors are generated after word segmentation and semantic embedding are conducted on user input information, and the text vectors are input into an intention recognizer and an entity extractor; matching the user motion intention and the key entity information with a preset slot position; when missing slot position information exists in the preset slot position, corresponding prompt information is generated according to the missing slot position information and fed back to a user; updating the key entity information based on the user supplementary information; searching an external knowledge base by taking the motion intention of the user and the updated key entity information as keywords; based on the retrieval result of the external knowledge base, the motion intention of the user and the updated key entity information, utilizing a prompt project to construct a standardized template; and a standardized template is utilized to guide the large language model, and a personalized exercise scheme meeting user requirements is generated. According to the method, the personalized exercise scheme meeting user requirements can be generated.

Description

Technical Field

[0001] The present application relates to the technical field of data processing, and particularly to a method, system and medium for generating a conversational exercise plan based on a large language model. Background Art

[0002] In the fast-paced modern life, people pay more and more attention to physical health, and exercise has become an effective means to enhance physical fitness, such as running, brisk walking, swimming, cardio training, etc. Due to individual differences, not all exercises are suitable for everyone. Therefore, formulating a personalized exercise plan for each person is crucial for improving physical fitness and enhancing immunity.

[0003] However, traditional exercise plans are usually manually customized according to an individual's living habits and status. This manual customization method is not only time-consuming and laborious, but also difficult to meet the diverse needs of users and cannot effectively achieve the customization of personalized exercise plans. Although there are also related technologies that use fixed templates and large language models to achieve this, most of these technologies still require adjusting the large language model, which also increases computing resources and is not conducive to the generation of personalized exercise plans. Summary of the Invention

[0004] The purpose of the present application is to provide a method, system and medium for generating a conversational exercise plan based on a large language model, which can relatively easily generate personalized exercise plans that meet the needs of users.

[0005] To achieve the above purpose, the present application provides the following solutions:

[0006] In a first aspect, the present application provides a method for generating a conversational exercise plan based on a large language model. The method for generating a conversational exercise plan based on a large language model includes:

[0007] Obtain user input information; the user input information is information related to exercise;

[0008] After performing word segmentation and semantic embedding on the user input information, generate a text vector and input it into an intent recognizer and an entity extractor; the intent recognizer is used to recognize the user's exercise intent according to the text vector; the entity extractor is used to extract key entity information according to the text vector;

[0009] Match the user's exercise intent and the key entity information with preset slots;

[0010] When there is missing slot information in the preset slots, generate corresponding prompt information according to the missing slot information and feedback it to the user;

[0011] Obtain user supplementary information and update the key entity information based on the user supplementary information;

[0012] Retrieve the external knowledge base using the user's exercise intention and the updated key entity information as keywords;

[0013] Based on the retrieval results of the external knowledge base, the user's exercise intention, and the updated key entity information, use prompt engineering to construct a standardized template;

[0014] Use the standardized template to guide the large language model to generate a personalized exercise plan that meets the user's needs.

[0015] In a second aspect, the present application also provides a computer system, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the method for generating a conversational exercise plan based on a large language model described in the first aspect.

[0016] In a third aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the method for generating a conversational exercise plan based on a large language model described in the first aspect.

[0017] According to the specific embodiments provided by the present application, the following technical effects are disclosed:

[0018] The present application establishes a mutual connection between the user and the system in a conversational manner, and uses a dynamic prediction slot checking mechanism to guide the user to supplement key information to ensure the integrity and personalization of the final exercise plan. At the same time, the present application also retrieves relevant information related to the user's exercise intention and key entity information through an external knowledge base, providing a scientific basis for the recommendation of the exercise plan. In addition, the present application uses prompt engineering to construct a standardized template and uses this to guide the large language model to generate an exercise plan, without the need to adjust the large language model again, making it easier to generate a personalized exercise plan that meets the user's needs. Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a flowchart of the method for generating a conversational exercise plan based on a large language model provided by an embodiment of the present application;

[0021] Figure 2 It is an execution block diagram of a personalized exercise plan provided by an embodiment of the present application;

[0022] Figure 3 This is the internal structure diagram of the computer system provided by the embodiment of the present application. Detailed implementation manners

[0023] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0024] The existing motion plan generation methods mainly have the following problems: First, they lack comprehensiveness and flexibility, relying more on fixed rules or templates and being difficult to handle diverse user needs. For example, some applications may only select exercise items based on basic information such as age and gender, without fully considering individual differences such as health status and living habits. Second, they have poor adaptability, mostly using static datasets for training and unable to be updated in real time to reflect the latest scientific research results and personal changes, which may lead to inaccurate recommended motion plans or unsuitability for the user's current physical conditions. Third, they have insufficient accuracy, failing to make full use of context information and performing poorly in understanding user intentions and entities in complex contexts, thus affecting the accuracy of the recommended results. Finally, the information is fragmented, and there is no effective mechanism to integrate knowledge from different sources, making the generated content possibly scattered and incoherent and unable to form a complete and coordinated motion plan.

[0025] However, the present application aims to use artificial intelligence technologies, especially natural language processing and deep learning models, to analyze the user's fitness goals, physical conditions, and other relevant information by means of a conversation with the user, and provide customized exercise suggestions for the user based on this data. By combining advanced technologies such as pre-trained language models, conditional random fields, large language models, and information-enhanced retrieval, efficient intent recognition and entity extraction can be achieved, making full use of the capabilities of large language models to ensure the high personalization and scientific nature of the motion plan.

[0026] The purpose of the present application is to provide a conversational motion plan generation method, system, and medium based on a large language model, which can relatively easily generate personalized motion plans that meet user needs.

[0027] To make the above objects, features, and advantages of the present application more obvious and understandable, the present application will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0028] Embodiment 1

[0029] AsFigure 1 and Figure 2 As shown in Figure 2 , this embodiment provides a method for generating a conversational exercise plan based on a large language model. The method for generating a conversational exercise plan based on a large language model includes:

[0030] Step S1: Obtain user input information.

[0031] In this embodiment, the user input information is information related to exercise. For example, the user inputs "I want to exercise to make my body look stronger".

[0032] Step S2: After tokenizing and semantically embedding the user input information, generate a text vector and input it into an intent recognizer and an entity extractor.

[0033] In this embodiment, the intent recognizer is used to recognize the user's exercise intent according to the text vector. Its structure is BERT+MLP, and its training sample example is:

[0034] {Feature (text vector): "I want to exercise to make my body look stronger"

[0035] Label (user exercise intent): "Muscle gain"}

[0036] In this embodiment, the entity extractor is used to extract key entity information according to the text vector. These key entity information (also known as annotation categories) can include age, hobbies, health status, exercise venue, exercise experience, height, weight, etc. In fact, it is to make annotations on the key information in the text in the BIO format (B category: mark the beginning of the entity of this category; I category: mark the middle part of the entity of this category; O: mark the words that do not belong to any category). The structure of this entity extractor is BERT+CRF, and its training sample example is:

[0037] {

[0038] "text": "I am 30 years old, male, 175 cm tall, 70 kg in weight, like strength training, and mainly in the gym",

[0039] "tags":

[0040] {"token": "I", "tag": "O"},

[0041] {"token": "this year", "tag": "O"},

[0042] {"token": "30 years old", "tag": "B-age"},

[0043] {"token": ",", "tag": "O"},

[0044] {"token": "Yes", "tag": "O"},

[0045] {"token": "Male", "tag": "B-Gender"},

[0046] {"token": ",", "tag": "O"},

[0047] {"token": "Height", "tag": "B-Height"},

[0048] {"token": "175 cm", "tag": "I-Height"},

[0049] {"token": ",", "tag": "O"},

[0050] {"token": "Weight", "tag": "B-Weight"},

[0051] {"token": "70 kg", "tag": "I-Weight"},

[0052] {"token": ",", "tag": "O"},

[0053] {"token": "likes", "tag": "O"},

[0054] {"token": "strength training", "tag": "B-Hobby"},

[0055] {"token": ",", "tag": "O"},

[0056] {"token": "mainly", "tag": "B-Sports Venue"},

[0057] {"token": "at", "tag": "I-Sports Venue"},

[0058] {"token": "the gym", "tag": "I-Sports Venue"},

[0059] {"token": ".", "tag": "O"}

[0061] }

[0062] Changing the format gives:

[0063] {

[0064] "Age": "30 years old",

[0065] ​"Gender": "Male",

[0066] "Height": "175 cm",

[0067] "Weight": "70 kg",

[0068] "Hobbies": ["Strength training"],

[0069] "Sports venue": "Gym"

[0070] }

[0071] Step S3: Match the user's sports intention and key entity information with the preset slots.

[0072] In this embodiment, the preset slots are actually the label sets in all the training samples of the intention recognizer and the entity extractor. At least two preset slots are set, and the slot information includes the user's sports intention, as well as age, height, weight, hobbies, health status, sports venue, and sports experience, etc.

[0073] In the actual matching process, the output results of the intention recognizer and the entity extractor can be converted into the following format:

[0074] {"Fitness intention": "Muscle gain", "Age": "30 years old", "Gender": "Male", "Height": "175 cm", "Weight": "70 kg", "Hobbies": "Strength training"}

[0075] Traverse this result and compare it with all the slot information (i.e., including all the labels) to check which values are missing or do not exist.

[0076] Step S4: When there is missing slot information in the preset slots, generate corresponding prompt information according to the missing slot information and feedback it to the user.

[0077] In this embodiment, for the missing slot information (such as age, height, etc.), the system generates targeted prompts and feedbacks them to the user. For example: "Please supplement your age and height information to generate a more personalized plan." After receiving this prompt information, the user will input supplementary information again.

[0078] Step S5: Obtain the user's supplementary information and update the key entity information based on the user's supplementary information.

[0079] In this embodiment, after obtaining the user's supplementary information, it is necessary to tokenize and semantically embed the user's supplementary information to generate a text vector, and then input it into the entity extractor again; when the key entity information determined based on the user's supplementary information successfully matches the slot information missing from the preset slot, the key entity information determined based on the user input information and the user's supplementary information is used as the updated key entity information.

[0080] Step S6: Use the user's exercise intention and the updated key entity information as keywords to retrieve the external knowledge base.

[0081] In this embodiment, the construction of the external knowledge base is to input a large number of exercise-related books and materials into existing large language models (such as GPT, Qwen, etc.), and use the large language model to assist in extracting entities (such as actions, muscles, body parts, etc.) and relationships (such as the association between actions and body parts), and construct knowledge triples (entity 1 - relationship - entity 2) to form a structured knowledge graph. At the same time, in order to maintain the timeliness and accuracy of the external knowledge base, it is also necessary to continuously add new data and knowledge related to exercise to achieve dynamic updates of the external knowledge base. In addition, the storage form of the knowledge graph in the external knowledge base is as follows:

[0082] Entity: "Running"

[0083] - Attributes:

[0084] * Type: Aerobic exercise

[0085] * Intensity level: Medium

[0086] * Goal: Fat loss

[0087] * Venue: Outdoor / Gym

[0088] - Relationships:

[0089] * "Suitable for" -> "Fat loss"

[0090] * "Requires" -> "Outdoor"

[0091] * "Includes" -> "Main muscle group: Lower limbs"

[0092] * "Precautions" -> "Knee injury"

[0093] Entity: "Knee injury"

[0094] - Attributes:

[0095] * Taboo exercises: Jumping, Squatting

[0096] * Recommended exercises: Swimming, Low-impact training

[0097] - Relationships:

[0098] *"Precautions" -> "Running"

[0099] *"Recommendation" -> "Swimming"

[0100] Step S7: Based on the retrieval results of the external knowledge base, the user's exercise intention, and the updated key entity information, use prompt engineering to construct a standardized template.

[0101] In this embodiment, the standardized template includes the following core parts: 1. Descriptions of the user's basic information, user goals, user exercise intention, and entity information. 2. The retrieval results of the external knowledge base, that is, all the content retrieved using the user's exercise intention and the updated key entity information as keywords. 3. The requirements for the final output exercise plan, including training plans, diet and recovery arrangements, safety precautions, etc. This is just a basic requirement for an exercise plan and has not yet generated the final exercise plan. It is used to guide what kind of exercise plan the large language model needs to generate and what specific requirements there are for the generated exercise plan. For example: requiring the large language model to give specific exercise actions, how many times to do a set, how many sets in total, what precautions there are for this action, and what kind of diet suggestions to match. A specific example of the standardized template is as follows:

[0102] User Basic Information

[0103] Fitness intention: {Fitness intention} # For example: fat loss, muscle gain, improving endurance

[0104] Age: {Age}

[0105] Gender: {Gender}

[0106] Height: {Height}

[0107] Weight: {Weight}

[0108] Hobbies: {Hobbies} # For example: strength training, running, yoga

[0109] Exercise venue: {Exercise venue} # For example: gym, home, outdoors

[0110] Exercise experience: {Exercise experience} # For example: beginner, intermediate, experienced fitness enthusiast

[0111] Health status: {Health status} # For example: knee injury, avoid high-impact exercises

[0112] ---

[0113] Knowledge Graph Query Results

[0114] Recommended exercise type: {Recommended exercise} # For example: strength training + aerobic training

[0115] Suitable target: {Suitable target} # For example: weight loss

[0116] Recommended exercises:

[0117] {Recommended exercises} # For example: deadlift, squat, push-up, HIIT training, etc.

[0118] Training intensity: {Training intensity} # For example: low, medium, high

[0119] Precautions: {Precautions} # For example: avoid heavy weight training, control heart rate

[0120] ---

[0121] Requirements for the exercise plan

[0122] a. Generate an exercise plan considering the above information

[0123] b. The exercise plan should include links such as warm-up, core training plan, relaxation, nutrition, and precautions

[0124] c. The plan should explain the specific exercises, how to do them, how many sets, how many reps or how long for each set, rest intervals, and training frequency.

[0125] Step S8: Use a standardized template to guide the large language model to generate a personalized exercise plan that meets the user's needs.

[0126] In this embodiment, the large language model can be selected from GPT, Qwen, etc. After the large language model generates the exercise plan, the system will feedback the generated exercise plan to the user and display the result in text. The user can further adjust the exercise plan or put forward new requirements, and the system will return to step S1 according to the new user input information to re-identify the user's exercise intention and adjust the plan.

[0127] An example of the personalized exercise plan finally output by the large language model is as follows:

[0128] **Personalized exercise plan**

[0129] # **Training plan**

[0130] Goal: {Exercise goal}

[0131] Training days: {Training frequency}

[0132] Venue: {Training venue}

[0133] Experience level: {Experience level}

[0134] # **Training content**

[0135] 1 **Warm-up (5 - 10 minutes)**

[0136] - Purpose: Increase heart rate, activate muscles, and prevent sports injuries

[0137] - Movements:

[0138] -{Warm-up movements} # For example: dynamic stretching, high knees, jumping jacks

[0139] - Sets: {Warm-up sets} # For example: 2 sets

[0140] - Repetitions: {Warm-up repetitions} # For example: 30 seconds / set

[0141] 2 **Core training program**

[0142] - **Training items**:

[0143] -{Training item 1}: {Training movement 1}

[0144] - **Target muscle groups**: {Target muscle groups 1}

[0145] - **Sets**: {Sets 1}

[0146] - **Repetitions**: {Repetitions 1}

[0147] - **Rest time**: {Rest time 1}

[0148] -{Training item 2}: {Training movement 2}

[0149] - **Target muscle groups**: {Target muscle groups 2}

[0150] - **Sets**: {Sets 2}

[0151] - **Repetitions**: {Repetitions 2}

[0152] - **Rest time**: {Rest time 2}

[0153] -{Training item 3}: {Training movement 3}

[0154] - **Target muscle groups**: {Target muscle groups 3}

[0155] - **Sets**: {Sets 3}

[0156] - **Repetitions**: {Repetitions 3}

[0157] - **Rest time**: {Rest time 3}

[0158] 3 **Aerobic training (optional)**

[0159] - Recommended Items: {Aerobic Exercises} # For example, easy running, cycling - Training Time: {Aerobic Time} # For example, 20 - 30 minutes - Intensity: {Aerobic Intensity} # For example, moderate intensity, heart rate 130 - 150 bpm 4** Relaxation & Recovery (5 - 10 minutes)**

[0160] - Purpose: Alleviate fatigue and promote muscle recovery

[0161] - Actions:

[0162] -{Relaxation Actions} # For example, static stretching, foam rolling

[0163] - Sets: {Relaxation Sets}

[0164] - Duration of Each Action: {Relaxation Time}

[0165] **Personalized Nutrition Recommendations**

[0166] # **Total Calorie Intake**

[0167] - Goal: {Exercise Goal} # For example, fat loss, muscle gain

[0168] - Recommended Total Calories per Day: {Calorie Intake} kcal

[0169] - Carbohydrates: {Carbohydrate Ratio}% (Recommended: 50 - 60%)

[0170] - Protein: {Protein Ratio}% (Recommended: 20 - 30%)

[0171] - Fat: {Fat Ratio}% (Recommended: 10 - 20%)

[0172] # **Recommended Foods**

[0173] - **High - Protein Foods**: {Protein Foods} # For example, chicken breast, beef, fish, protein powder

[0174] - **Healthy Fats**: {Fatty Foods} # For example, avocado, nuts, olive oil

[0175] - **High - Quality Carbs**: {Carbohydrate Foods} # For example, brown rice, oats, sweet potatoes

[0176] # **Water Intake**

[0177] - Recommended Daily Water Intake: {Water Intake} L

[0178] - During Exercise: Drink {Exercise Water Intake} ml every 20 minutes

[0179] # **Supplement Recommendations (Optional)**

[0180] -{Supplements}#For example, caffeine before exercise, protein powder after exercise, creatine, etc.

[0181] -Target population:{Target population for supplements}#For example, those who need to improve endurance

[0182] **Precautions**

[0183] {Precautions}

[0184] Example 2

[0185] This example provides a computer device, which can be a server or a terminal, and its internal structure diagram can be as shown in Figure 3 the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store user input information and user supplement information. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements the above-mentioned method for generating a conversational exercise plan based on a large language model.

[0186] Those skilled in the art can understand that Figure 3 the structure shown in is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have a different component layout.

[0187] Example 3

[0188] This example provides a computer-readable storage medium storing a computer program, which when executed by a processor implements the steps in the above method examples.

[0189] In an exemplary embodiment, a computer program product is provided, including a computer program, which when executed by a processor implements the steps in the above method examples.

[0190] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0191] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0192] The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., and are not limited thereto. The processors involved in the embodiments provided in this application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., and are not limited thereto.

[0193] All actions of obtaining signals, information, or data in this application are carried out on the premise of complying with the corresponding data protection regulations and policies of the country where it is located and with the authorization given by the owner of the corresponding device.

[0194] In summary, the present application mainly has the following advantages:

[0195] (1) Natural language interaction: Achieve user-friendly natural language input parsing through large language models, and accurately identify users' exercise goals and needs.

[0196] (2) Dynamic information supplementation: Design a dynamic slot check mechanism to guide users to supplement key information and ensure the integrity and personalization of the recommended solutions.

[0197] (3) Knowledge graph support: Construct and real-time update the knowledge graph in the field of sports-related, providing a scientific basis for solution recommendation.

[0198] (4) Solution generation and adjustment: Based on prompt engineering, generate a sports solution prompt template and support users to adjust and optimize dynamically in real time.

[0199] (5) Wide applicability: Applicable to fitness beginners and non-professional users, providing flexible, efficient and scientific sports solution recommendation services.

[0200] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other.

[0201] Specific examples are used in this article to elaborate on the principles and implementation methods of the present application. The descriptions of the above embodiments are only used to help understand the method and its core idea of the present application; at the same time, for those of ordinary skill in the art, according to the idea of the present application, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A method for generating a conversational exercise plan based on a large language model, characterized in that The method for generating a conversational exercise plan based on a large language model includes: Obtaining user input information; the user input information is information related to exercise; Performing word segmentation and semantic embedding on the user input information to generate a text vector, and inputting it into an intent recognizer and an entity extractor; the intent recognizer is used to recognize the user's exercise intent according to the text vector; the entity extractor is used to extract key entity information according to the text vector; Matching the user's exercise intent and the key entity information with a preset slot; When there is missing slot information in the preset slot, generating corresponding prompt information according to the missing slot information and feeding it back to the user; Obtaining user supplementary information and updating the key entity information based on the user supplementary information; Using the user's exercise intent and the updated key entity information as keywords to retrieve an external knowledge base; Based on the retrieval result of the external knowledge base, the user's exercise intent, and the updated key entity information, constructing a standardized template using prompt engineering; Using the standardized template to guide the large language model to generate a personalized exercise plan that meets the user's needs.

2. The method for generating a conversational exercise plan based on a large language model according to claim 1, wherein, The structure of the intent recognizer is BERT+MLP.

3. The method for generating a conversational exercise plan based on a large language model according to claim 1, wherein, The structure of the entity extractor is BERT+CRF.

4. The method for generating a conversational exercise plan based on a large language model according to claim 1, wherein, At least 2 preset slots are set.

5. The method for generating a conversational exercise plan based on a large language model according to claim 4, wherein The slot information includes the user's exercise intent and any one of age, height, weight, hobbies, health status, exercise venue, and exercise experience.

6. The method for generating a conversational exercise plan based on a large language model according to claim 1, wherein Updating the key entity information based on the user supplementary information specifically includes: Performing word segmentation and semantic embedding on the user supplementary information to generate a text vector, and inputting it into the entity extractor; When the key entity information determined based on the user supplementary information matches the missing slot information in the preset slot, using the key entity information determined based on the user input information and the user supplementary information as the updated key entity information.

7. The method for generating a conversational exercise plan based on a large language model according to claim 1, wherein The standardized template includes at least the user's basic information, the retrieval result of the external knowledge base, and the requirements for the exercise plan.

8. The method for generating a conversational motion plan based on a large language model according to claim 1, wherein: The large language model is at least any one of GPT and Qwen.

9. A computer system, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the method for generating a conversational exercise plan based on a large language model according to any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the method for generating a conversational exercise plan based on a large language model according to any one of claims 1-8.

Images