Cooking method based on large cookery ai model and ai agent system, and intelligent cooking apparatus for same

Abstract

Disclosed are a cooking method based on a large cookery AI model and an AI agent system, and an intelligent cooking apparatus for same. The intelligent cooking apparatus is provided with a user health condition sensing apparatus, an information collection module, a touch screen, a storage module, a processor, and a wireless communication module. The information collection module is used to receive requirement information and a health goal of a user, and the user health condition sensing apparatus is used to acquire user physical health condition information and push to the AI agent system the requirement information of the user, the health goal set by the user, the user physical health condition information, and historical diet record data of the user. The AI agent system outputs a health management plan matching the health goal set by the user and executes a cooking sub-task plan in the health management plan by means of a usage tool module of the AI agent system. The present invention applies the AI agent system and the large cookery AI model to the intelligent cooking apparatus, thereby achieving the health management goal set by the user and improving the physical health condition of the user.

Description

A cooking method based on a cooking AI large model and an AI agent system and an intelligent cooking device thereof TECHNICAL FIELD

[0001] The present application relates to a cooking method and device thereof, in particular to a cooking method based on a cooking AI large model and an AI agent system and an intelligent cooking device thereof. BACKGROUND

[0002] With the development of economy and society, people's life rhythm is getting faster and faster, the cost of life is increasing, the competition pressure is increasing, and the time planning is getting more and more nervous, so that sleep is reduced, exercise is less and diet is irregular, which is easy to appear fatigue, drowsiness, body aches, insomnia, emaciation, function decline, function disorder and other sub-health states.

[0003] The existing intelligent cooking device can realize automatic cooking, save time cost and ensure fresh and delicious food, but it cannot make reasonable diet planning for user personalization to improve the user's sub-health state. SUMMARY

[0004] In order to solve the above technical problems, the present application provides a cooking method based on a cooking AI large model and an AI agent system, applied to an intelligent cooking device, the intelligent cooking device is provided with a user health condition sensing device, an information collection module, a touch screen, a storage module, a processor and a wireless communication module, the information collection module, the user health condition sensing device, the touch screen, the storage module, the wireless communication module and the processor are electrically connected, the information collection module is used for receiving user demand information input by user through voice, video, text, picture or 3D model format and user set health target, the user health condition sensing device is used for obtaining user body health condition information, the storage module is used for recording user's historical diet record data, the AI agent system is provided with a memory module, a planning module, an execution module and a tool module, the user demand information, the user set health target, the user body health condition information and the user's historical diet record data are pushed to the AI agent system, the AI agent system outputs a health management plan meeting the user set health target, including the following three steps:

[0005] Firstly, the AI agent system receives the user set health target and sends the user set health target to the planning module;

[0006] The second step is that the planning module formulates a user health diet management task according to the cooking knowledge base of the memory module, the historical diet record data of the user and the health goal set by the user, the user health diet management task includes mutual collocation between multiple food materials in a specific period, and the user health diet management task is decomposed into multiple cooking sub-tasks, wherein each cooking sub-task includes a recipe for any one meal in a period.

[0007] The third step is that the AI agent system formulates a corresponding cooking sub-task plan according to each cooking sub-task and sends it to the user mobile terminal or the touch screen, and the user confirms or modifies it by operating the user mobile terminal or the touch screen, wherein the cooking sub-task modified by the user returns to the AI agent system for the second step of re-planning of the health management plan conforming to the health goal set by the user, and the health management plan conforming to the health goal set by the user is generated until the health management plan conforming to the health goal set by the user is generated, and each cooking sub-task plan includes an execution date and an execution time.

[0008] The AI agent system adjusts the cooking sub-task plan and executes the corresponding cooking sub-task by calling the application tool through the use tool module, wherein the application tool includes a calendar tool and a time tool, the use tool module calls the calendar tool and the time tool to set the execution date and the execution time of the third step of the health management plan output by the AI agent system conforming to the health goal set by the user, the AI agent system calls a cooking AI large model to train the recipe of the second step of the health management plan output by the AI agent system conforming to the health goal set by the user into a new cooking method, the new cooking method includes cooking operation instructions and cooking parameters, and the cooking AI large model performs inference operation on the new cooking method according to the execution date and the execution time and sends it to the user mobile terminal or the touch screen through the wireless communication module after the inference operation is completed, and the user confirms the new cooking method by operating the user mobile terminal or the touch screen, and the intelligent cooking device cooks food according to the new cooking method and prompts the user to complete the cooking operation instructions that need the user to cooperate.

[0009] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, the AI agent system creates a cooking knowledge base by collecting cooking information in the cooking industry, training the cooking information, obtaining cooking data and storing the cooking data in the memory module to create the cooking knowledge base, and the application tool further includes an external data source, and the use tool module calls the external data source to collect new cooking information in the cooking industry to update the cooking knowledge base during the execution of the health management plan.

[0010] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, the AI agent system outputs each of the cooking sub-task plans of the third step of the health management plan conforming to the health goal set by the user, which further includes a recipe plan, the recipe plan including preparation of cooking utensils, food materials, seasonings, and execution sequence between each of the cooking sub-tasks, the application tool further including a food material purchase platform APP and a cooking environment detection device, the use tool module calling the food material purchase platform APP to purchase food materials and seasonings, or reminding the user to purchase through the calendar tool and the time tool, and the use tool module calling the cooking environment detection device to detect the cooking utensils in the cooking environment.

[0011] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, the application tool further includes a query modification tool, and the user can query and modify the unexecuted cooking sub-task plan through the query modification tool during the execution of the health management plan, the use tool module feeds back the modified sub-task plan to the AI agent system, and the planning module modifies and adjusts the health management plan until the health goal set by the user is completed.

[0012] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, the application tool further includes an evaluation tool, and the use tool module calls the evaluation tool to perform self-evaluation and summary of experience of the completed cooking sub-tasks based on the health goal set by the user, so as to guide the user to adjust the health management plan until the health goal set by the user is completed.

[0013] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, the application tool further includes a calculation tool, the recipe plan limiting the amount of food materials and the amount of seasonings, the use tool module calling the calculation tool to record and calculate the intake amount of each nutritional element and energy of each of the completed cooking sub-tasks according to the amount of food materials and the amount of seasonings in the food material plan, and feeding back to the AI agent system, and the AI agent system modifying and adjusting the health management plan according to the information fed back by the calculation tool until the health goal set by the user is completed.

[0014] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, during the execution of the health management plan process, the use tool module calls the user health condition sensing device to detect the user's physical health condition information in real time and feeds back to the AI agent system. The AI agent system modifies and adjusts the health management plan according to the real-time detection of the user's physical health condition information fed back by the user health condition sensing device, aiming at the health goal set by the user, until the health goal set by the user is completed.

[0015] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, the user health diet management task includes health regulation mode recipes, treatment mode recipes or free arrangement mode recipes.

[0016] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, the intelligent cooking device includes the environment sensing device for obtaining the environment information of the user, the health regulation mode recipe is a health management plan output by the AI agent system based on the cooking knowledge base of the memory module according to the environment information of the user, the user's physical health condition information and the user's historical diet record data; the treatment mode recipe is a health management plan output by the AI agent system based on the cooking knowledge base of the memory module according to the pathology input by the user through the information collection module and the user's historical diet record data; the free arrangement mode recipe is a health management plan output by the AI agent system based on the cooking knowledge base of the memory module according to the cooking recipe input by the user through the information collection module, the user's physical health condition information and the user's historical diet record data.

[0017] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, the health goal set by the user includes weight target, heart rate index, blood pressure index, blood oxygen concentration, bone density, body composition index, seasonal health diet target, regional health diet target or population health diet target.

[0018] As an improvement of the cooking method based on the cooking AI large model and the AI agent system of the present application, the cooking AI large model generates a new cooking method including two steps: constructing a cooking AI large model and calling a cooking AI large model, and the construction of a cooking AI large model includes the following six steps:

[0019] The first step is to collect healthy cooking data: collect all information about food healthy cooking methods presented in the form of voice, video, text, pictures or 3D model, including all food cooking method information, body health status and food healthy cooking method relationship information, weather environment status and food healthy cooking method relationship information;

[0020] The second step is to preprocess the healthy cooking data: preprocess all the collected information about food healthy cooking methods to ensure the integrity and availability of the information, including converting different formats of information into text, and editing the text information according to certain formats to facilitate the subsequent training of AI large model;

[0021] The third step is to select AI large models applicable to cooking: select third-party AI large models at home and abroad, and measure them with accuracy, response speed and diversity indicators;

[0022] The fourth step is to train the cooking AI large model: after the second step, the healthy cooking data set is sorted out, and then the healthy cooking data set is fine-tuned with the third-party AI large model. After training, a cooking AI large model with all related data of food healthy cooking method is generated;

[0023] The fifth step is to verify and test the cooking AI large model: the cooking AI large model generated in the fourth step is tested for specific task effect evaluation. If the evaluation effect does not pass, the first step, second step, third step and fourth step are repeated for retraining until the effect evaluation passes, and the cooking AI large model is generated and deployed in the storage module of the cloud platform or intelligent cooking device;

[0024] The sixth step is to deploy and maintain the cooking AI large model: deploy the newly generated cooking AI large model to the storage module of the cloud platform or intelligent cooking device, and continuously maintain and update it, regularly update the data to ensure the timeliness and accuracy of the data;

[0025] The calling cooking AI large model specifically includes the following steps:

[0026] The first step is to assemble the query statement: the collected user body health status information and user environment information are transmitted to the cooking AI large model;

[0027] The second step is to perform inference operation by the cooking AI large model: the query statement is transmitted to the cooking AI large model, and the inference operation is performed by the cooking AI large model, which includes the following three steps:

[0028] 1) Understand the input: distributed semantic parsing, which first receives a text sequence and converts it into a word vector. This process is based on the distributed semantic hypothesis that the meaning of a word is determined by its use in context;

[0029] 2) Parameter association: context-focus chaining, inputting these word vectors into the Transformer Encoder to generate context representation;

[0030] 3) Generate Answers: Generative probabilistic modeling: The model initializes the Transformer's decoder and feeds the encoder output and the current output sequence into the decoder. The decoder generates a probability distribution for the next word and selects the word with the highest probability or another set probability distribution as the output. This word is then appended to the output sequence.

[0031] 4) Select the most appropriate answer: Dynamic word string evolution, repeat the above steps, adding new words to the output sequence each time, until a complete output sequence is generated;

[0032] Step 3. The cooking AI big model returns the result: After the cooking AI big model completes the inference operation, it returns the information related to the new healthy cooking method or new recipe and the operation instructions of the new healthy cooking method or new recipe, including providing the user with content in text, picture, audio, video or 3D model format.

[0033] The present invention provides an intelligent cooking device, including a user health status sensing device, an information collection module, a touch screen, a storage module, a processor and a wireless communication module. The information collection module, the user health status sensing device, the touch screen, the storage module, the wireless communication module are electrically connected to the processor. The information collection module is used to receive user demand information and health goals set by the user. The user health status sensing device is used to obtain user physical health information. The storage module can record the user's historical diet record data and health management plan. The processor is connected to the cloud platform and the user's mobile terminal through the wireless communication module. The processor is used to execute the new cooking method generated by the cooking AI large model in claim 1 and prompt the user to perform cooking operation instructions.

[0034] As an improvement of the intelligent cooking device of the present invention, the intelligent cooking device is provided with cooking operation stations for stir-frying, sautéing, deep-frying, cooking, frying, sticking, roasting, braising, stewing, steaming, blanching, boiling, stewing, sautéing, mixing, marinating, roasting, braising, freezing, pulling out silk, honey-glazing, smoking, rolling, sliding or baking.

[0035] As an improvement to the intelligent cooking device of the present invention, the cooking operation stations of stir-frying, sautéing, deep-frying, cooking, frying, sticking, roasting, braising, stewing, steaming, blanching, boiling, stewing, sautéing, mixing, marinating, roasting, braising, freezing, candied food, honey-glazed food, smoking, rolling, sliding or baking are provided with corresponding operation detection feedback systems, which are used to detect whether the cooking operations performed by the user meet the requirements of the new cooking method of the cooking AI large model.

[0036] As an improvement to the intelligent cooking device of the present invention, a human-computer interaction module is provided, and the human-computer interaction module is provided with a voice recognition device. The human-computer interaction module is used for information interaction and intelligent control between the intelligent cooking device and the user. The intelligent control includes the user confirming cooking parameters and cooking operation instructions through voice or on the touch screen.

[0037] The present invention provides a cooking method and intelligent cooking device based on a cooking AI big model and an AI agent system, and has the following beneficial effects: by pushing the user demand information, the health goals set by the user, the user's physical health information, and the user's historical dietary record data to the AI ​​agent system, the AI ​​agent system outputs a health management plan that meets the health goals set by the user, so as to formulate a personalized health management plan for the user; the health management plan is reasonably decomposed into multiple cooking subtasks through the planning module, facilitating the user's planned execution and improving execution efficiency; each cooking subtask includes a recipe for any diet within a cycle, and the user can confirm or modify the cooking subtask plan by operating the user mobile terminal or the touch screen; the user can adjust the ingredients of the cooking subtask plan to their favorite ingredients according to their dietary preferences, avoiding user resistance and improving the execution of the health management plan; the user can be reminded by the use tool module to confirm the execution of the relevant subtask plan to supervise the smooth progress of the health management plan, and the cooking AI big model is called to train the recipe into a new cooking method and send it to the intelligent cooking device to cook food according to the new cooking method, realizing intelligent cooking and saving time and cost. The agent system can develop personalized health management plans for users and supervise users to implement health management plans on time and efficiently. At the same time, it can ensure that the smart device cooks food that suits the user's taste to achieve the user's set health management goals and improve the user's physical health. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] FIG1 is a workflow diagram of a preferred embodiment of a cooking method and an intelligent cooking device thereof based on a cooking AI big model and an AI agent system according to the present invention.

[0039] Figure 2 is a workflow diagram of other embodiment one of the cooking method and intelligent cooking device based on the cooking AI large model and AI agent system of the present application.

[0040] Figure 3 is a workflow diagram of other embodiment two of the cooking method and intelligent cooking device based on the cooking AI large model and AI agent system of the present application.

[0041] Figure 4 is a workflow diagram of other embodiment three of the cooking method and intelligent cooking device based on the cooking AI large model and AI agent system of the present application.

[0042] Figure 5 is a workflow diagram of other embodiment four of the cooking method and intelligent cooking device based on the cooking AI large model and AI agent system of the present application.

[0043] Figure 6 is a workflow diagram of other embodiment five of the cooking method and intelligent cooking device based on the cooking AI large model and AI agent system of the present application.

[0044] Figure 7 is a workflow diagram of the AI agent system of the preferred embodiment of the cooking method and intelligent cooking device based on the cooking AI large model and AI agent system of the present application.

[0045] Figure 8 is a flowchart of the AI agent system outputting a health management plan that meets the health goal set by the user of the preferred embodiment of the cooking method and intelligent cooking device based on the cooking AI large model and AI agent system of the present application. DETAILED DESCRIPTION

[0046] In the following, the present application will be further described in conjunction with the accompanying drawings 1-8 and the specific embodiments and other embodiments, it should be noted that the technical features described below can be combined in any combination to form new embodiments without conflict.

[0047] In a preferred embodiment, with reference to FIGS. 1, 7, 8, the present application provides a cooking method based on a cooking AI large model 4 and an AI agent system 5, applied to an intelligent cooking device, the intelligent cooking device is provided with a user health condition sensing device 11, an information collection module 12, a touch screen 14, a storage module 17, a processor 15 and a wireless communication module 16, the information collection module 12, the user health condition sensing device 11, the touch screen 14, the storage module 17, the wireless communication module 16 and the processor 15 are electrically connected, the information collection module 12 is used to receive user 33 demand information and health goals set by the user 33 in the form of voice, video, text, picture or 3D model format, the user health condition sensing device 11 is used to obtain user 33 body health condition information, the storage module 17 is used to record user 33 historical diet record data, the AI agent system 5 is provided with a memory module 53, a planning module 51, an execution module 52 and a tool module 54, the user demand information, the health goals set by the user 33, the user 33 body health condition information and the user 33 historical diet record data are pushed to the AI agent system 5, and the AI agent system 5 outputs a health management plan meeting the health goals set by the user 33, including the following three steps:

[0048] 601, the AI agent system 5 receives the health goals set by the user 33 and sends the health goals set by the user 33 to the planning module 51;

[0049] 602, the planning module 51 formulates a user health diet management task according to the cooking knowledge base of the memory module 53, the historical diet record data of the user 33 and the health goals set by the user 33, the user health diet management task includes mutual matching between a plurality of food materials in a specific period, and each cooking subtask includes a recipe for any one meal in a period;

[0050] 603, the AI agent system 5 formulates a corresponding cooking subtask plan according to each cooking subtask and sends it to the user mobile terminal 31 or the touch screen 14, and the user 33 confirms or modifies it by operating the user mobile terminal 31 or the touch screen 14, wherein the cooking subtask modified by the user 33 returns to the AI agent system 5 to output a second step of re-planning of the health management plan meeting the health goals set by the user 33, until the health management plan meeting the health goals set by the user 33 is generated, and each cooking subtask plan includes an execution date and an execution time;

[0051] The AI agent system 5 deploys the cooking sub-task plan and executes the corresponding cooking sub-task by calling the application tool 541 through the use tool module 54, wherein the application tool 541 includes a calendar tool 5412 and a time tool 5413, and the use tool module 54 calls the calendar tool 5412 and the time tool 5413 to set the execution date and execution time of the third step of the health management plan output by the AI agent system 5 to meet the health goal set by the user 33; the AI agent system 5 calls the cooking AI large model 4 to train the recipe of the second step of the health management plan output by the AI agent system 5 to meet the health goal set by the user 33 into a new cooking method, the new cooking method including cooking operation instructions and cooking parameters, and the cooking AI large model 4 sends the new cooking method to the user mobile terminal 31 or the touch screen 14 through the wireless communication module 16 after reasoning and operation of the new cooking method according to the execution date and execution time, and the user 33 confirms the new cooking method by manipulating the user mobile terminal 31 or the touch screen 14, and the intelligent cooking device cooks food according to the new cooking method and prompts the user 33 to complete the cooking operation instructions that need the user 33 to cooperate.The planning module includes various general AI large models, such as M6 of Alibaba, Wudao 2.0 of Zhiyuan Research Institute, Hunyuan of Tencent, Wenxin of Baidu, Zidong·Taichu of the Automation Institute of the Chinese Academy of Sciences, ChatGLM3 of ZhiPu Huzhang, BaiChuan of BaiChuan Intelligence, BookSheng of Shanghai Artificial Intelligence Experiment, XingHuo of the National University of Defense Technology, InternLM of Simgot, DouBao of ChunTian ZhiYun (TikTok), TongYi QianWen of Alibaba Cloud, and PangGu of Huawei, etc. The memory module 53 can be a knowledge base corresponding to the general AI large model. For example, the health goal input by the user 33 is “lose 8 pounds in a month”, the AI agent system 5 sends “lose 8 pounds in a month” and historical diet record data to the AI agent system 5, the general AI large model of the planning module 51 is InternLM, which is first trained as a health management AI large model with professional cooking knowledge and professional fitness and weight loss nutrition guidance knowledge. The health management AI large model of the planning module 51 analyzes and formulates a user health diet management task for the health goal “lose 8 pounds in a month” and historical diet record data according to the cooking knowledge base and fitness and weight loss knowledge base of the health management AI large model, and the health management AI large model reasons and decomposes the user health diet management task into multiple cooking sub-tasks. One of the sub-tasks includes “eat chicken breast for dinner”, and the health management AI large model formulates a corresponding cooking sub-task plan according to each cooking sub-task, such as a plan for “eat chicken breast for dinner”, which includes cold chicken breast. The plan for “eat chicken breast for dinner” is sent to the user mobile terminal 31, and the user 33 does not like cold chicken breast, modifies the plan for “eat chicken breast for dinner” to include grilled chicken breast, and feeds back to the health management AI large model for re-planning. Finally, the health management plan for “lose 8 pounds in a month” is output.

[0052] In a preferred embodiment, the AI agent system 5 creates a cooking knowledge base by collecting cooking information in the cooking industry, training the cooking information, obtaining cooking data, and storing the cooking data in the memory module 53 to create the cooking knowledge base. The application tool 541 also includes an external data source 5411. During the execution of the health management plan process, the use tool module 54 calls the external data source 5411 to collect new cooking information in the cooking industry to update the cooking knowledge base. The external data source 5411 includes online search engines, applications, web browsers or communication platforms such as Baidu, Sogou, 360 Search, Google China, Youtube, Notion, Xiaohongshu, and private professional cooking knowledge bases.

[0053] In the preferred embodiment, the AI agent system 5 outputs each of the cooking subtask plans of the third step of the health management plan that meets the health goal set by the user 33 further includes a recipe plan, the recipe plan includes preparation of cooking utensils, food materials, seasonings, and execution sequence between each of the cooking subtasks, the application tool 541 further includes a food material purchase platform APP 5414 and a cooking environment detection device 5418, the use tool module 54 calls the food material purchase platform APP 5414 to purchase food materials and seasonings, or reminds the user 33 to purchase through the calendar tool 5412 and the time tool 5413, and the use tool module 54 calls the cooking environment detection device 5418 to detect the cooking utensils in the cooking environment, and sets the calendar tool 5412 and the time tool 5413 to supervise the user 33 to execute the health management plan.

[0054] In the preferred embodiment, the application tool 541 further includes a query modification tool 5417, during the execution of the health management plan, the user 33 can query and modify the unexecuted cooking subtask plan through the query modification tool 5417, the use tool module 54 feeds back the modified subtask plan to the AI agent system 5, and the planning module 51 modifies and adjusts the health management plan until the health goal set by the user 33 is achieved, and the query modification tool 5417 is set to ensure that the subtask can be executed smoothly. For example, when the cooking environment detection device 5418 does not detect the steamer required by the cooking utensils of the cooking subtask plan, the cooking utensils of the cooking subtask plan are modified to the existing oven in the query modification tool 5417, and feedback is given to the AI agent system 5, the planning module 51 modifies and adjusts the health management plan until the health goal set by the user 33 is achieved; for another example, the user queries the cooking subtask plan to be executed through the query modification tool 5417 and finds that there is an undesirable food material, celery, then the celery in the cooking subtask plan is modified to be replaced by agaric in the query modification tool 5417, and feedback is given to the AI agent system 5, the planning module 51 modifies and adjusts the health management plan until the health goal set by the user 33 is achieved.

[0055] In the preferred embodiment, the application tool 541 further comprises an evaluation tool 5416, and the use tool module 54 calls the evaluation tool 5416 to perform self-evaluation and summary experience of the completed cooking sub-tasks based on the health goal set by the user 33, so as to guide the user 33 to adjust the health management plan until the health goal set by the user 33 is achieved; by setting the evaluation tool 5416, the effect of the health management plan implementation can be checked at any time, the efficiency of achieving the health goal is improved, and the time cost and cooking cost of the user are avoided. For example, the health goal set by the user is “to lose 8 pounds in a month”, and the health management plan formulated for “to lose 8 pounds in a month” is divided into four cycles, one week as a cycle, and the weight loss target of the first cycle is to lose 2-3 pounds. After the first cycle of the health management plan is implemented, the weight loss effect achieved is 1.5 pounds, which does not reach the expected effect, and compared with the health goal “to lose 8 pounds in a month”, 6.5 pounds still needs to be lost in the remaining three cycles. The evaluation tool 5416 feeds back the gap between the weight loss effect and the health goal to the AI agent system 5, and the planning module 51 modifies and adjusts the health management plan until the health goal set by the user 33 is achieved.

[0056] In the preferred embodiment, the application tool 541 further comprises a calculation tool 5415, and the amount of food material and the amount of seasoning in the recipe plan are limited. The use tool module 54 calls the calculation tool 5415 to record and calculate the intake amount of each nutritional element and energy of each completed cooking sub-task according to the amount of food material and the amount of seasoning limited in the food material plan, and feeds back to the AI agent system 5. The AI agent system 5 modifies and adjusts the health management plan according to the information fed back by the calculation tool 5415 until the health goal set by the user 33 is achieved; for example, the user does not perform the cooking sub-task according to the amount of food material and the amount of seasoning limited in the recipe plan of the cooking sub-task, and the intake amount of each nutritional element and energy exceeds the limit of the cooking sub-task. The calculation tool calculates the gap between the intake amount of each nutritional element and energy of the completed sub-task and the total intake amount of each nutritional element and energy limited in the health management plan, feeds back to the AI agent system 5, and the AI agent system 5 modifies and adjusts the health management plan until the health goal set by the user 33 is achieved.

[0057] In the preferred embodiment, during the execution of the health management plan by the application tool 541, the use tool module 54 calls the user health condition sensing device 11 to detect the user's 33 physical health condition information in real time and feeds back to the AI agent system 5. The AI agent system 5 modifies the health management plan according to the real-time detection of the user's 33 physical health condition information fed back by the user health condition sensing device 11, aiming at the health goal set by the user 33, until the health goal set by the user 33 is completed; by detecting the user's 33 physical health condition in real time, the user's physical condition is monitored in real time to avoid abnormality of the user's body. For example, the user health condition sensing device 11 is a blood pressure sensor, and it is detected that the user's blood pressure is lower than 90 / 60 mmHg (the normal blood pressure range of adults is 100-139 / 60-89 mmHg), so the intake of salt, high protein, iron and vitamin and other elements required by the human body needs to be increased, and fed back to the AI agent system 5. The planning module 51 modifies the health management plan until the health goal set by the user 33 is completed. For another example, the user health condition sensing device 11 is an electronic scale, and the user's set health goal is "to lose 8 pounds in a month". The health management plan formulated for "to lose 8 pounds in a month" is divided into four periods, one week as a period, and the first period target is to lose 2-3 pounds. After the implementation of the first period according to the health management plan, it is measured by the electronic scale that 4 pounds are lost, which is too fast and may harm health. The user health condition sensing device 11 feeds back the difference between the weight loss effect and the health goal to the AI agent system 5, and the planning module 51 modifies the health management plan until the health goal set by the user 33 is completed.

[0058] In a preferred embodiment, the user 33 health diet management task includes a health regimen recipe, a treatment regimen recipe, or a free arrangement regimen recipe. The user 33 physical health information includes facial features, obesity level, heart rate data, blood pressure index, blood oxygen concentration, or hoarseness level, the user health condition sensing device 11 includes an AI camera, a heart rate sensor, a blood pressure sensor, a blood oxygen sensor, a body temperature sensor, or a voice recognition device, which are electrically connected to the processor; the smart cooking device includes the environment sensing device, which is used to obtain the user 33 environment information, including the solar term, the season, the regional location, the weather condition, or the temperature and humidity condition, the environment sensing device includes a weather sensor, a temperature sensor, a humidity sensor, a GPS device, a light sensor, or a gas sensor, which are electrically connected to the processor. The health regimen recipe is a health management plan output by the AI agent system 5 based on the user 33 environment information, user physical health information, and user historical diet record data based on the cooking knowledge base of the memory module; for example, through the humidity sensor to determine that the air in the user 33 location is dry, and in terms of diet, it is suitable to moisten the lung and remove dry heat, and clear heat and detoxify, through the blood oxygen sensor to find that the user 33 blood oxygen saturation is high, and in terms of diet, it is suitable to be light and low in salt, the AI agent system 5 outputs a health management plan based on the cooking knowledge base of the memory module according to the user environment information fed back by the humidity sensor, the user physical health information fed back by the blood oxygen sensor, and the user historical diet record data, wherein the health management plan includes recipes for moistening the lung and removing dry heat, clearing heat and detoxifying, light, and low in salt, such as silver ear snow pear soup, stir-fried winter melon, and seaweed beef soup.

[0059] The treatment mode recipe is a health management plan output by the AI agent system 5 based on the cooking knowledge base of the memory module according to the pathology and historical dietary record data of the user 33 input through the information collection module; for example, the pathology input by the user 33 through the information collection module is postoperative rehabilitation, and the AI agent system outputs a health management plan based on the cooking knowledge base of the memory module according to the pathology and historical dietary record data of the user 33 input through the information collection module, wherein the health management plan includes light, easy to digest and absorb, and nutritious recipes that help wound healing, such as boiled beef slices, fish head tofu soup, steamed lean meat, etc.; for another example, the pathology input by the user 33 through the information collection module is postpartum conditioning, and the information collection module pushes the information of "postpartum conditioning" input by the user 33 to the AI agent system 5, and the AI agent system 5 outputs a health management plan based on the cooking knowledge base of the memory module according to the pathology and historical dietary record data of the user 33 input through the information collection module, wherein the health management plan includes easy-to-digest, nutritious, and iron-rich recipes to help blood replenishment, such as steamed fish, mustard pig liver soup, and beef porridge. The free arrangement mode recipe is a health management plan output by the AI agent system 5 based on the cooking knowledge base of the memory module according to the cooking recipe, user health information and historical dietary record data of the user 33 input through the information collection module, and the cooking recipe input by the user 33 through the information collection module is the user's favorite recipe, for example, the cooking recipe input by the user 33 through the information collection module is roast ribs, and the body temperature sensor detects that the user's health condition is low fever, the AI agent system 5 analyzes that the user is not suitable for eating roast ribs according to the cooking recipe and the historical dietary record data of the user 33 input through the information collection module based on the cooking knowledge base of the memory module, and needs to eat light, easy-to-digest, and high-moisture food, and the recipe of the health management plan output by the AI agent system 5 is changed to include bitter gourd rib soup.

[0060] In the preferred embodiment, the health goal set by the user 33 includes a weight goal, a heart rate indicator, a blood pressure indicator, a blood oxygen concentration, a bone density, a body composition indicator, a seasonal health diet goal, a regional health diet goal, or a population health diet goal; the seasonal health diet goal includes seasonal vegetables and some recipes combined with the local climate environment, for example, the recipes in Guangdong in winter include stewed radish with mutton, cordyceps sinensis and chicken, yam fish fillet, and braised lotus root with pork ribs, etc., and the recipes in Guangdong in summer include spicy and sour potato shreds, hand-torn Chinese cabbage, stir-fried zucchini, and sweet and sour pork ribs, etc. The regional health diet goal includes local dietary customs, for example, the recipes in coastal areas are inclined to aquatic products, such as shrimps, scallops, fish, crabs, sea tangle, dragon spine, stone flower, kelp, and laver, etc.; the recipes in Sichuan are inclined to be spicy, for example, maopo tofu, kung pao chicken, and steamed fish, etc.; the recipes in Hunan are inclined to be salty and spicy, for example, chopped chili fish head, Mao's braised pork, spicy chicken, and chili fried pork, etc. The population health diet goal includes special groups, for example, the recipes for the fitness group need a reasonable proportion of carbohydrates, proteins, and fats, for example, lemon grilled salmon, pasta with tomato and bass, and steak with potatoes, etc.; the recipes for the child group need to pay attention to the balanced nutrition of proteins, fats, carbohydrates, minerals, vitamins, and water, for example, steamed hairtail, lean meat with bean sprouts, steamed hairtail, chicken with carrots, and duck with beans, etc.; the recipes for the elderly group need to be low in fat, low in cholesterol, rich in nutrition, and easy to digest, for example, stewed pork ribs with potatoes, and steamed pork ribs with garlic, etc.; the recipes for the pregnant group need to be balanced in nutrition, high in calcium, and high in protein, for example, corn and pork rib soup, yam and chicken soup, and old duck soup, etc.; the recipes for the postoperative rehabilitation group need to be light, easy to digest, and rich in nutrition to help wound healing, for example, boiled beef slices, fish head and tofu soup, and steamed lean meat, etc.; the recipes for the postpartum recovery group need to be easy to digest, rich in nutrition, and high in iron to help blood recovery, for example, steamed fish, mustard and pork liver soup, and beef porridge, etc.

[0061] In the preferred embodiment, the cooking AI large model 4 generates a new cooking method including two steps: constructing a cooking AI large model 41 and calling a cooking AI large model 42, and the constructing a cooking AI large model 41 includes the following six steps:

[0062] 101, Collecting healthy cooking data: collecting all information about food health cooking methods presented in the form of voice, video, text, picture, or 3D model, including all food cooking method information, body health status and food health cooking method relationship information, weather environment status and food health cooking method relationship information;

[0063] 102、Preprocessing healthy cooking data: all the information collected about the healthy cooking methods of food is preprocessed to ensure the integrity and availability of the information, including converting information of different formats into text and editing the text information according to a certain format to facilitate the subsequent training of the AI large model;

[0064] 103、Selecting AI large models applicable to cooking: selecting third-party AI large models at home and abroad, and measuring them with accuracy, response speed, and diversity indicators;

[0065] 104、Training cooking AI large model 4: After the healthy cooking data set is sorted in the second step, the healthy cooking data set is fine-tuned and trained by the third-party AI large model. After the training is completed, the cooking AI large model 4 with all the relevant data of the healthy cooking method of food is generated;

[0066] 105、Verification test cooking AI large model 4: The cooking AI large model 4 generated in the fourth step is detected and evaluated for specific tasks. If the evaluation effect does not pass, the steps of the first step, the second step, the third step, and the fourth step are repeated for retraining until the effect evaluation passes. The cooking AI large model 4 is generated and deployed in the storage module 17 of the cloud platform or the intelligent cooking device;

[0067] 106、Deployment and maintenance of cooking AI large model 4: The newly generated cooking AI large model 4 is deployed into the storage module 17 of the cloud platform or the intelligent cooking device, and is continuously maintained and updated, and the data is regularly updated to ensure the timeliness and accuracy of the data;

[0068] The calling cooking AI large model 42 specifically includes the following steps:

[0069] 201、Assemble query statement; the collected user 33 health information and the environment information of the user 33 are transmitted to the cooking AI large model 4;

[0070] 202、Cooking AI large model 4 performs inference operation: the query statement is transmitted to the cooking AI large model 4, and the cooking AI large model 4 performs inference operation, which includes the following three steps:

[0071] 202-1、Understanding input: distributed semantic parsing, which first receives a text sequence and converts it into a word vector. This process is based on the distributed semantic hypothesis that the meaning of a word is determined by its use in context;

[0072] 202-2、Parameter association: context focus chain, input these word vectors into the Encoder of the Transformer to generate context representation;

[0073] 202-3, generating answers: generative probability modeling, model initialization Transformer Decoder part, and input the output of the Encoder and the current output sequence into the Decoder, the Decoder will generate the probability distribution of the next word, select the word with the maximum probability or other set probability distribution as the output, this word will be added to the output sequence;

[0074] 202-4, selecting the most suitable answer: dynamic word string evolution, repeat the above steps, add new words to the output sequence each time, until a complete output sequence is generated;

[0075] 203, cooking AI large model 4 returns the result: after the cooking AI large model 4 inference operation is completed, the new healthy cooking method or new recipe related information and the new healthy cooking method or new recipe operation instruction are returned, including providing the user 33 with text, picture, audio, video or 3D model format content.

[0076] Referring to FIG. 1, the present application provides an intelligent cooking device, comprising a user health condition sensing device 11, an information collection module 12, a touch screen 14, a storage module 17, a processor 15 and a wireless communication module 16, the information collection module 12, the user health condition sensing device 11, the touch screen 14, the storage module 17, the wireless communication module 16 and the processor 15 are electrically connected, the information collection module 12 is used to receive user 33 demand information and user 33 set health target, the user health condition sensing device 11 is used to obtain user 33 body health condition information, the storage module 17 can record user 33 historical diet record data, health management plan, the processor 15 is connected between the cloud platform and the user mobile terminal 31 through the wireless communication module 16, the processor 15 is used to execute the new cooking method generated by the cooking AI large model 4 in claim 1 and prompt user 33 to perform cooking operation instruction. The intelligent cooking device is provided with frying 26, frying 25, frying 22, stewing 21, steaming 23 and cooking 24 cooking operation stations, the frying 26 cooking operation station is a cooking machine, the frying 25 cooking operation station is an air fryer, the frying 22 cooking operation station is a frying and baking machine, an electromagnetic oven, a microwave oven or an oven, the stewing 21 cooking operation station is a slow cooker, an electric stewing pot or an electric stewing cup, the steaming 15 cooking operation station is a steaming and baking oven, an electric steaming pot, an electric rice cooker or an electric pressure cooker, and the cooking 16 cooking operation station is an electric cooking pot. One or more intelligent cooking devices use one or more cooking functions of frying, air frying, frying, stewing, steaming and cooking to cook food materials. The processor 15 can set the working parameters of the cooking operation stations of the intelligent cooking device and control the corresponding cooking operation stations to execute the related cooking operation instructions, for example, if the new cooking method includes frying pork slices, after the new cooking method including frying pork slices is sent to the intelligent cooking device, the processor sets the working parameters of the frying 14 cooking operation station of the intelligent cooking device and controls the frying 14 cooking operation station to standby, start, put oil, put pork slices, stir-fry, stop and clean.

[0077] Referring to FIG. 1, in a preferred embodiment, the cooking method based on the cooking AI large model and the AI agent system and the intelligent cooking device thereof of the present application, the cooking AI large model 4 and the AI agent system 5 are arranged on the cloud platform 22.

[0078] Referring to FIG. 6, in other embodiment five, the cooking method based on the cooking AI large model and the AI agent system and the intelligent cooking device thereof of the present application, the cooking AI large model 4 and the AI agent system 5 are built-in in the storage module 17.

[0079] Referring to FIG. 1, in the preferred embodiment, the intelligent cooking device is provided with cooking operation stations of frying 26, frying 25, frying 22, stewing 21, steaming 23 and boiling 24.

[0080] Referring to FIG. 2, in another embodiment one, the intelligent cooking device is provided with cooking operation stations of frying 25, frying 22, stewing 21, steaming 23 and boiling 24.

[0081] Referring to FIG. 3, in another embodiment two, the intelligent cooking device is provided with cooking operation stations of frying 22, stewing 21, steaming 23 and boiling 24.

[0082] Referring to FIG. 4, in another embodiment three, the intelligent cooking device is provided with cooking operation stations of frying 22 and stewing 21.

[0083] Referring to FIG. 5, in another embodiment four, the intelligent cooking device is provided with a cooking operation station of stewing 21.

[0084] Referring to FIG. 1, in the preferred embodiment, the cooking operation stations of frying 26, frying 25, frying 22, stewing 21, steaming 23 and boiling 24 are provided with corresponding operation detection feedback systems 18 for detecting whether the cooking operation performed by the user 33 meets the requirements of the new cooking method of the cooking AI large model 4. The operation detection feedback system 18 includes camera, infrared detection, radar detection, magnetic detection, weight detection and other detection devices for detecting whether the cooking operation performed by the user in the intelligent cooking device meets the requirements of the new cooking method of the cooking AI large model 4. For example, at the cooking operation station of stewing 21, the intelligent cooking device requires the user to add an appropriate amount of water, but the user forgets to add water. At this time, the operation detection feedback system 18 detects through the camera and feeds back the information to the processor 15 of the intelligent cooking device. The processor 15 controls the cooking operation station to pause the next operation and sends information to the user 33 for correction. When the user 33 corrects, the operation detection feedback system 18 detects through the camera and feeds back the information to the processor 15. The processor 15 controls the cooking operation station to perform the next operation. The above is only an example, and in addition, infrared detection, radar detection, magnetic detection, weight detection or a combination thereof can be used for detection.

[0085] Referring to FIG. 1, in the preferred embodiment, the intelligent cooking device is provided with a human-computer interaction module 13, which is provided with a voice recognition device, and is used for information interaction and intelligent control between the intelligent cooking device and the user 33, including confirmation of cooking parameters and cooking operation instructions by the user 33 through voice or the touch screen. The human-computer interaction module 13 also includes an image input unit and a camera device, and is electrically connected with the processor 15, and is used for information interaction between the intelligent cooking device and the user. The user 33 can input user demand information and user-set health goals in the form of voice, video, text, picture or 3D model through the voice recognition device, the image input unit or the camera device. The human-computer interaction module 13 sends the user demand information and the user-set health goals to the processor 15. The processor 15 pushes the user demand information, the user-set health goals, the user's physical health condition information and the user's historical diet record data to the AI agent system through the wireless communication module. The AI agent system outputs a health management plan that meets the user-set health goals. The AI agent system calls relevant application tools through the use tool module to execute a cooking sub-task plan of the health management plan. The cooking AI large model 4 is called to analyze and train the recipe of the cooking sub-task plan to generate a new cooking method, which includes cooking operation instructions and cooking parameters. After the new cooking method is completed by inference operation of the cooking AI large model 4, it is sent to the human-computer interaction module 13 (or the user mobile terminal 31) and the intelligent cooking device through the wireless communication module 16. The processor 15 determines the cooking operation station of the intelligent cooking device according to the new cooking method, sets relevant parameters for the cooking operation station, and then controls the cooking operation station to cook. During the cooking process, when some steps need to be operated by the user 33, the processor 15 sends information to the user 33 through the human-computer interaction module 13 (or the user mobile terminal 31) in the form of image, voice or text to prompt the user 33 to complete the cooking operation instructions that need to be operated by the user 33.

[0086] In other embodiments, the intelligent cooking device is one or a combination of a wok, an air fryer, an induction cooker, a microwave oven, an oven, a steam oven, an electric rice cooker, a steak machine, a barbecue machine, a pot rice machine, an electric pressure cooker, a steam cooker, a multifunctional food processor, an electric pot, a bread maker, a chef machine, a steam pot machine, an electric ceramic stove, a cooking pot, a frying and baking machine, a slow cooker, an electric stew pot, an electric stew cup, an integrated stove, or a frying pan, and one or more intelligent cooking devices include one or more cooking functions of frying, exploding, rolling, frying, cooking, frying, sticking, burning, stewing, stewing, steaming, boiling, cooking, cooking, frying, frying, frying, marinating, baking, stewing, freezing, pulling silk, honey, smoking, rolling, sliding, or baking to cook food; the cooking operation station of frying, exploding, rolling, frying, cooking, frying, sticking, burning, stewing, stewing, steaming, boiling, cooking, cooking, frying, frying, frying, marinating, baking, stewing, freezing, pulling silk, honey, smoking, rolling, sliding, or baking is provided with a corresponding operation detection feedback system, which is used to detect whether the cooking operation performed by the user meets the requirements of the new cooking method of the cooking AI large model.

[0087] In other embodiments, the intelligent cooking device includes multiple cooking functions including frying, air frying, baking, frying, stewing, stewing, steaming, boiling, or baking, such as an intelligent cooking device that integrates air frying, baking, frying, steaming, and other functions, or an intelligent cooking device that integrates frying, air frying, stewing, stewing, boiling, and other functions.

[0088] The cooking method based on the cooking AI large model and the AI agent system and the intelligent cooking device have the beneficial effects that: the user demand information, the user set health target, the user physical health condition information and the user historical diet record data are pushed to the AI agent system, the AI agent system outputs a health management plan meeting the user set health target, so as to formulate a personalized health management plan for the user, the health management plan is reasonably decomposed into a plurality of cooking sub-tasks through the planning module, the user can execute the plan, the execution efficiency is improved, each cooking sub-task includes a recipe of any one-time diet in a period, and the user can confirm or modify the cooking sub-task plan through the user mobile terminal or the touch screen, the user can adjust the cooking sub-task plan to the favorite food material according to the own diet preference, the user's resistance psychology is avoided, the health management plan execution degree is improved, the user can be reminded to confirm the execution of the related sub-task plan through the use tool module, the health management plan is supervised to proceed smoothly, the cooking AI large model is called to train the recipe into a new cooking method and send the intelligent cooking device to cook food according to the new cooking method, the intelligent cooking is realized, the time cost is saved, the AI agent system is set, the personalized health management plan can be formulated for the user, the user can be supervised to execute the health management plan on time and efficiently, the intelligent device can cook food meeting the user's taste, so that the user's set health management target is realized, and the user's physical health condition is improved.

[0089] The above has made a detailed description of the present application, the above is only the preferred embodiment of the present application, which cannot limit the scope of the present application, that is, any equivalent change and modification made within the scope of the present application shall still fall within the scope of the present application.

Claims

1. A cooking method based on a cooking AI large model and an AI agent system, characterized by, The application is applied to an intelligent cooking device provided with a user health condition sensing device, an information collection module, a touch screen, a storage module, a processor and a wireless communication module. The information collection module, the user health condition sensing device, the touch screen, the storage module, the wireless communication module and the processor are electrically connected. The information collection module is used for receiving user demand information and health target set by a user in a voice, video, text, picture or 3D model format. The user health condition sensing device is used for obtaining user body health condition information. The storage module is used for recording historical dietary record data of the user. An AI agent system is provided with a memory module, a planning module, an execution module and a tool module. The user demand information, the health target set by the user, the user body health condition information and the historical dietary record data of the user are pushed to the AI agent system. The AI agent system outputs a health management plan meeting the health target set by the user, including the following three steps. Firstly, the AI agent system receives the health target set by the user and sends the health target to the planning module. Secondly, the planning module formulates a user health dietary management task according to a cooking knowledge base of the memory module, the historical dietary record data of the user and the health target set by the user. The user health dietary management task includes mutual matching among various food materials in a specific period. Each cooking subtask includes a recipe for any one-time diet in a period. Thirdly, the AI agent system formulates a corresponding cooking subtask plan according to each cooking subtask and sends it to a user mobile terminal or the touch screen. The user confirms or modifies it by operating the user mobile terminal or the touch screen. The cooking subtask modified by the user returns to the second step of the AI agent system to output a health management plan meeting the health target set by the user for re-planning until the health management plan meeting the health target set by the user is generated. Each cooking subtask plan includes an execution date and an execution time. The AI agent system deploys the cooking sub-task plan and calls application tools to execute corresponding cooking sub-tasks, wherein the application tools include a calendar tool and a time tool, the use tool module calls the calendar tool and the time tool to set the execution date and time of the third step of the health management plan output by the AI agent system to meet the health goal set by the user; the AI agent system calls a cooking AI large model to train the recipe of the second step of the health management plan output by the AI agent system to meet the health goal set by the user into a new cooking method, the new cooking method includes cooking operation instructions and cooking parameters, and the cooking AI large model sends the new cooking method to the user mobile terminal or the touch screen through the wireless communication module after reasoning and operation of the new cooking method according to the execution date and time, and the user confirms the new cooking method by operating the user mobile terminal or the touch screen, and the intelligent cooking device cooks according to the new cooking method and prompts the user to complete the cooking operation instructions that need the user to cooperate. 2.The cooking method based on a cooking AI large model and an AI agent system according to claim 1, characterized in that: The AI agent system creates a cooking knowledge base by collecting cooking information in the cooking industry, training the cooking information, obtaining cooking data and storing the cooking data in the memory module to create the cooking knowledge base, and the application tools also include external data sources, and the use tool module calls the external data sources to collect new cooking information in the cooking industry to update the cooking knowledge base during the execution of the health management plan. 3.The cooking method based on the cooking AI large model and the AI agent system of claim 1, wherein: The AI agent system outputs the third step of the health management plan that meets the health goal set by the user, and each cooking sub-task plan also includes a recipe plan, the recipe plan includes preparation of cooking utensils, food materials, seasonings and execution sequence between each cooking sub-task, and the application tools also include a food material purchase platform APP and a cooking environment detection device, the use tool module calls the food material purchase platform APP to purchase food materials and seasonings, or reminds the user to purchase through the calendar tool and the time tool, and the use tool module calls the cooking environment detection device to detect the cooking utensils in the cooking environment. 4.The cooking method based on a cooking AI large model and an AI agent system according to claim 1, characterized in that: The application tools also include a query modification tool, and the user can query and modify the unexecuted cooking sub-task plan through the query modification tool during the execution of the health management plan, the use tool module feeds back the modified sub-task plan to the AI agent system, and the planning module modifies and adjusts the health management plan until the health goal set by the user is completed. 5.The cooking method based on a cooking AI large model and an AI agent system according to claim 1, characterized in that: The application tools also include an evaluation tool, and the use tool module calls the evaluation tool to self-evaluate and summarize experience of the completed cooking sub-tasks based on the health goal set by the user, so as to guide the user to adjust the health management plan until the health goal set by the user is completed. 6.The cooking method based on a cooking AI large model and an AI agent system according to claim 1, characterized in that: The application tool also includes a calculation tool that defines the amount of food and the amount of seasoning in the recipe plan, and the use tool module calls the calculation tool to record and calculate the amount of each nutrient and energy intake of each cooking sub-task completed according to the amount of food and the amount of seasoning defined in the food plan, and feeds back to the AI agent system, which modifies and adjusts the health management plan according to the information fed back by the calculation tool until the user's health goal is achieved. 7.The cooking method based on the cooking AI large model and the AI agent system of claim 1, wherein: During the execution of the health management plan by the application tool, the use tool module calls the user health condition sensing device to detect the user's physical health condition information in real time and feeds back to the AI agent system, which modifies and adjusts the health management plan according to the real-time detection of the user's physical health condition information fed back by the user health condition sensing device for the user's health goal until the user's health goal is achieved. 8.The cooking method based on the cooking AI large model and the AI agent system of claim 1, wherein: The user health diet management task includes a health regulation mode recipe, a treatment mode recipe, or a free arrangement mode recipe. 9.The cooking method based on the cooking AI large model and the AI agent system of claim 1, wherein: The intelligent cooking device includes the environment sensing device for obtaining the user's environment information, the health regulation mode recipe is a health management plan output by the AI agent system based on the cooking knowledge base of the memory module according to the user's environment information, user's physical health condition information, and user's historical diet record data; the treatment mode recipe is a health management plan output by the AI agent system based on the cooking knowledge base of the memory module according to the user's pathology input through the information collection module and the user's historical diet record data; the free arrangement mode recipe is a health management plan output by the AI agent system based on the cooking knowledge base of the memory module according to the user's cooking recipe input through the information collection module, user's physical health condition information, and user's historical diet record data. 10.The cooking method based on the cooking AI large model and the AI agent system of claim 1, wherein: The user's health goal includes weight target, heart rate index, blood pressure index, blood oxygen concentration, bone density, body composition index, seasonal health diet target, regional health diet target, or population health diet target. 11.The cooking method based on the cooking AI large model and the AI agent system of claim 1, wherein: The cooking AI large model generates a new cooking method including two steps: building a cooking AI large model and calling a cooking AI large model, the building a cooking AI large model includes the following six steps: First, collect health cooking data: collect all information about food health cooking methods presented in the form of voice, video, text, picture or 3D model, including all food cooking method information, body health condition and food health cooking method relationship information, weather environment condition and food health cooking method relationship information; The second step is to preprocess the healthy cooking data: all the collected information about the healthy cooking methods of food is preprocessed to ensure the integrity and availability of the information, including converting different formats of information into text and editing the text information according to a certain format to facilitate the subsequent training of the AI large model; The third step is to select an AI large model applicable to cooking: the third-party AI large models at home and abroad are selected, and the accuracy, response speed, and diversity indicators are used as the evaluation criteria; The fourth step is to train the cooking AI large model: after the healthy cooking data set is sorted out in the second step, the healthy cooking data set is fine-tuned and trained by the third-party AI large model, and the cooking AI large model with all the related data of the healthy cooking method of food is generated after the training is completed; The fifth step is to verify and test the cooking AI large model: the cooking AI large model generated in the fourth step is evaluated for specific task effects, and if the evaluation effect does not pass, the first step, the second step, the third step, and the fourth step are repeated for retraining until the effect evaluation passes, and the cooking AI large model is generated and deployed in the storage module of the cloud platform or the intelligent cooking device; The sixth step is to deploy and maintain the cooking AI large model: the newly generated cooking AI large model is deployed in the storage module of the cloud platform or the intelligent cooking device, and is continuously maintained and updated, and the data is updated regularly to ensure the timeliness and accuracy of the data; The calling of the cooking AI large model specifically includes the following steps: The first step is to assemble a query statement: the collected user's health information and the user's environment information are transmitted to the cooking AI large model together; The second step is for the cooking AI large model to perform inference operation: the query statement is transmitted to the cooking AI large model, and the cooking AI large model performs inference operation, which includes the following three steps: 1) Understanding input: distributed semantic parsing, which first receives a text sequence and converts it into a word vector, this process is based on the distributed semantic hypothesis that the meaning of a word is determined by its use in context; 2) Parameter association: context focus chain, input the word vector into the Encoder of the Transformer to generate a context representation; 3) Generate answer: generative probabilistic modeling, initialize the Decoder part of the Transformer, and input the output of the Encoder and the current output sequence into the Decoder, the Decoder will generate the probability distribution of the next word, select the word with the maximum probability or other set probability distribution as the output, and this word will be added to the output sequence; 4) Select the most appropriate answer: dynamic word string evolution, repeat the above steps, and add a new word to the output sequence each time until a complete output sequence is generated; The third step is for the cooking AI large model to return the result: after the inference operation of the cooking AI large model is completed, the cooking AI large model returns the related information of the new healthy cooking method or the new recipe and the operation instructions of the new healthy cooking method or the new recipe, including providing the user with text, pictures, audio, video, or 3D model format content.

12. A smart cooking device, characterized by, The application relates to a health management system, which comprises a user health condition sensing device, an information collection module, a touch screen, a storage module, a processor and a wireless communication module, wherein the information collection module, the user health condition sensing device, the touch screen, the storage module, the wireless communication module and the processor are electrically connected; the information collection module is used for receiving user demand information and user-set health targets; the user health condition sensing device is used for acquiring user physical health condition information; the storage module can record user historical diet record data; the processor is connected with a cloud platform and a user mobile terminal through the wireless communication module; and the processor is used for executing a new cooking method generated by the cooking AI large model in claim 1 and prompting a user to perform cooking operation instructions.

13. The smart cooking apparatus of claim 12, wherein: The cooking operation station is provided with a frying, exploding, rolling, frying, boiling, frying, pasting, burning, stewing, stewing, steaming, boiling, cooking, cooking, cooking, frying, mixing, marinating, baking, stewing, freezing, pulling silk, honey juice, smoking, rolling, sliding or baking operation station.

14. The smart cooking apparatus of claim 13, wherein: The cooking operation station is provided with a corresponding operation detection feedback system, which is used for detecting whether the cooking operation performed by the user meets the requirements of the new cooking method of the cooking AI large model.

15. The smart cooking apparatus of claim 12, wherein: The application further discloses a human-computer interaction module, which is provided with a voice recognition device and is used for information interaction and intelligent control between the intelligent cooking device and the user, wherein the intelligent control includes user confirmation of cooking parameters and cooking operation instructions through voice or the touch screen.

Images