Generative artificial intelligence-based banquet creative design system

Abstract

The invention discloses a generative artificial intelligence-based banquet creative design system, which comprises: a demand determination module, which is used for interacting with a user terminal, obtaining a user demand and generating a demand theme; the banquet creativity generation engine is used for determining and operating a proxy artificial intelligence assembly line and generating a banquet creativity design project; comprising a main control unit, a domain knowledge acquisition unit, a creativity generation unit and a integrating degree evaluation unit, each unit is embedded with a multi-level intelligent agent as an execution main body, and prompt information is defined as input of the execution main body; the brand domain knowledge vector library is used for providing domain knowledge; the banquet design domain knowledge vector library is used for providing banquet design knowledge; and the design output module is used for obtaining the banquet creative design project and forming a banquet creative design scheme. According to the technical scheme, multi-modal brand domain knowledge and banquet creative design can be deeply fused, and automatic generation of multi-modal information of a high-end banquet catering art scheme is achieved.

Description

Technical Field

[0001] This invention relates to the field of generative artificial intelligence applications, and more specifically, to a banquet creative design system based on generative artificial intelligence. Background Art

[0002] In the field of business banquet planning, the integration of menu design, ambiance setup, and brand theme is crucial for realizing the event's value. Currently, the industry generally relies on planning teams to manually design plans based on the experience of hotel chefs and the client's requirements. However, with societal progress, the demands for business banquets are becoming increasingly diverse, requiring the embodiment of brand concepts, promotion, and enhancement of brand image. Therefore, the dimensions to be considered in business banquet design are also increasing, such as event objectives, brand positioning, and event theme. Simple manual design can no longer meet these diverse design needs.

[0003] With the development of generative artificial intelligence technology, more and more designers are using generative AI agents to realize their creative designs. However, excellent banquet creative designs need to reflect the brand's professional needs in the design details, which in turn rely on the capabilities of the implementers (such as chefs). Therefore, in the process of determining the requirements proposed by designers to the generative AI agent, if it is based solely on personal experience, the process of generating design results will lack quantitative model support, and the correlation with the brand concept cannot be measured, thus preventing multiple rounds of optimization. Although information systems are currently used to realize creative designs, the feedback cycle for creative solutions is long, with high time and communication costs, frequent repetitions, and low efficiency. Chefs or banquet designers often rely solely on personal experience and find it difficult to deeply understand the brand's core concepts, professional knowledge in the brand domain, and cultural connotations, thus limiting the scope of creativity and making it difficult to quickly iterate and produce diverse and high-quality creative design solutions. Moreover, the banquet creative design process is often experience-driven, with a single decision-making process and a lack of multi-faceted and multi-dimensional deep collaborative mechanisms.

[0004] Therefore, a technical solution is needed to achieve cross-domain knowledge collaborative reasoning, and to automatically generate banquet creative designs through multiple rounds of interaction based on deep understanding of brand semantics and creative ideas for dishes, so as to improve the accuracy of matching creative solutions with customer needs, improve design quality, and enhance customer satisfaction with designed products. Summary of the Invention

[0005] To achieve the above objectives, this application provides a banquet creative design system based on generative artificial intelligence, comprising:

[0006] Requirements Determination Module: This module interacts with the user terminal to obtain user requirements, parses these requirements, and generates a requirements theme. The requirements theme refers to the service items proposed by the user to the banquet provider, and consists of multi-level attributes and attribute content. The multi-level attributes include first-level attributes and N-level attributes. The first-level attributes directly reflect the user's requirements. The requirements reflected by the first-level attributes are decomposed into dimensions that need to be confirmed during the design process, forming N-level attributes.

[0007] Banquet Creative Generation Engine: Used to identify and run a pipeline of agent-based artificial intelligence agents to generate banquet creative design projects based on the required theme; wherein, the banquet creative generation engine includes a main control unit, a domain knowledge acquisition unit, a creative generation unit, and a fit evaluation unit; each unit embeds multi-level intelligent agents as execution subjects, and defines the prompt information of each intelligent agent as the input of the execution subject;

[0008] Brand Domain Knowledge Vector Library: Used to provide domain knowledge related to user characteristics;

[0009] Banquet Design Knowledge Vector Library: Provides banquet design knowledge related to the actual capabilities of banquet implementers;

[0010] Design Output Module: Used to acquire banquet creative design projects output by the banquet creative generation engine and form banquet creative design schemes.

[0011] The main control unit embeds a main control agent, including: determining the corresponding input information as the main control prompt information, including defining the corresponding role information as the main control agent, the input information being the first-level attributes and attribute content of the demand topic, and the task requirement being to determine the brand domain knowledge scope for assisting in banquet design;

[0012] The master control agent is used to understand the task and plan the overall task based on the master control prompts, and outputs the brand domain knowledge scope of the auxiliary banquet design.

[0013] The domain knowledge acquisition unit is embedded with a domain knowledge vector library intelligent agent, including: determining the corresponding input information as brand domain prompt information, including defining the corresponding role information as brand knowledge agent, the input information being the brand domain knowledge scope of the auxiliary banquet design, and the task requirement being to retrieve and output brand knowledge;

[0014] The domain knowledge vector library agent, based on the brand domain prompts and combined with the brand domain knowledge vector library, performs in-depth analysis of user characteristics, executes dynamic retrieval, and outputs brand knowledge.

[0015] The creative generation unit embeds a creative intelligent agent, including: determining the corresponding input information as creative prompt information, including defining the corresponding role information as banquet design creative agent, the input information as brand knowledge and primary attributes and attribute content, and the task requirement as generating creative menu plans for the banquet;

[0016] The creative agent outputs creative dish proposals based on creative prompts and a knowledge vector library for the banquet design domain. The creative dish proposals consist of N-level attributes and attribute values ​​of the demand theme.

[0017] The matching degree evaluation unit embeds a matching degree evaluation intelligent agent, including: determining the corresponding input information as evaluation prompt information, including defining the role information as aesthetic matching degree Agent, the input information as the first-level attributes and attribute content of the creative dish plan and the demand theme, and the task requirement as generating aesthetic matching degree evaluation results;

[0018] The matching evaluation agent generates aesthetic matching evaluation information and provides feedback suggestions based on the evaluation prompts. The evaluation information includes evaluation results and optimization suggestions. The evaluation results include result confirmation and return for optimization.

[0019] When the evaluation result of the matching evaluation agent is confirmed, the creative dish plan is output by the banquet creative generation engine as a banquet creative design project. When the evaluation result is returned for optimization, the creative agent is called again, optimization suggestions are input, and the creative dish plan output by the creative agent is obtained again. Then, the aesthetic matching evaluation information is generated again through the matching evaluation agent.

[0020] The specific process of the agent-based AI pipeline includes: the output of the master AI agent is the input of the domain knowledge vector library AI agent; the output of the domain knowledge vector library AI agent is the input of the creative AI agent; the output of the creative AI agent is the input of the fit evaluation AI agent; and the evaluation result of the output of the fit evaluation AI agent determines whether to output the banquet creative design project.

[0021] Furthermore, the content of the first-level attributes and the categories of the N-level attributes are determined in the requirements determination module. In the banquet creative generation engine, the content of the N-level attributes is obtained by each agent-type artificial intelligence.

[0022] The banquet creative design system proposed in this invention can clearly distinguish the roles of intelligent agents such as planning, domain knowledge, creative generation, and banquet aesthetic theme fit assessment. Through structured collaboration, it significantly improves the overall quality and efficiency of creative generation, and systematically records the collaboration and decision-making process among multiple intelligent agents, making the reasons for banquet creative decisions clear and transparent, facilitating review and optimization. At the same time, it can deeply integrate multimodal brand domain knowledge with banquet creative design, and support an adaptive high-end business banquet creative iteration process with real-time feedback and targeted optimization from the client brand through multi-intelligent planning, multi-round reflection, and collaborative generation mechanisms, thereby ultimately achieving the automatic generation of multimodal information for high-end banquet catering art solutions. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the banquet creative design system provided in an embodiment of the present invention. Detailed Implementation

[0024] In real-world scenarios involving banquet creative design needs, the requirements of brand clients typically include objectives (e.g., brand promotion and enhancement of brand image), brand positioning (e.g., representing global luxury goods, emphasizing the combination of fashion and quality), event theme (e.g., a light pink and white theme conveying elegance and purity), and event scale and venue setup requirements (e.g., 100 attendees, a modern minimalist venue decorated with pink roses). When designers address these various needs, they usually also need to consider the style preferences of the hotel chefs and incorporate brand expertise. Based on this application background, this invention proposes a banquet creative design system that embeds an agent-based artificial intelligence algorithm framework, integrates brand expertise with a banquet design experience knowledge base, and achieves cross-domain knowledge collaborative reasoning and deep understanding of brand semantics. The system engages in multiple rounds of interaction during the reasoning process, supporting the automatic implementation of banquet creative designs.

[0025] The specific implementation of the present invention will now be described in detail with reference to the accompanying drawings.

[0026] The structure of the banquet creative design system proposed in this invention is as follows: Figure 1 As shown, it includes:

[0027] P100 Demand Determination Module: Used to interact with the user terminal, obtain the user's requirements, parse the user's requirements, and generate a demand theme; the demand theme refers to the service items proposed by the user to the banquet implementer, which consists of multi-level attributes and attribute content;

[0028] The multi-level attributes of a demand theme include first-level attributes and N-level attributes. First-level attributes can directly reflect the user's needs, including: theme style, service scale, and theme cultural connotation.

[0029] In the requirements definition module, the requirements reflected by the first-level attributes are decomposed and further subdivided into which dimensions need to be designed and confirmed in order to meet the user's requirements during the design process. These dimensions include the characteristics of the user (brand), i.e., user characteristics, as well as the characteristics that the banquet implementer needs to implement, i.e., banquet implementer characteristics.

[0030] Secondary attributes include the purpose of ingredient selection, cooking methods, dining experience, and interactive style. Further analysis can be conducted on secondary attributes to further subdivide them into tertiary attributes: the purpose of ingredient selection includes the color and shape of the ingredients, the dining experience includes the texture of the tableware and the restaurant lighting, and the interactive style includes the serving process and background music, thus forming N-level attributes.

[0031] For example, the theme style is: a light pink and white theme, conveying the concept of elegance and purity;

[0032] The event organizer needs to consider how to accommodate the client's requirements for a menu combining pink and white, and how to convey an elegant and pure concept. Through the definition of N-level attributes, designers are provided with dimensions to consider, such as the color of the dishes, the cuisine, and the interaction methods, reflecting the brand's domain details from the content of the N-level attributes.

[0033] Overall, the requirement determination module defines the content of the first-level attributes and the categories of the N-level attributes. In the subsequent banquet idea generation engine, each agent AI agent obtains the content of the N-level attributes and ultimately generates an effective banquet design.

[0034] P110 Banquet Creative Generation Engine: Used to identify and run a pipeline of agent-based artificial intelligence entities to generate banquet creative design projects based on demand themes.

[0035] The banquet creative generation engine introduces AgenticAI, which has capabilities such as autonomous planning, reflection, chain of thought, and retrieval augmented generation. It combines the banquet implementation knowledge base (such as menu structure, cooking methods, techniques, and aesthetic experience) with the brand domain knowledge base, and loads a multi-agent collaborative construction pipeline of agentic AI agents. The multi-agent agents can be implemented by the same or different large models.

[0036] Specifically, the banquet idea generation engine includes: P111 main control unit, P112 domain knowledge acquisition unit, P113 idea generation unit, and P114 fit evaluation unit. Multi-level intelligent agents are embedded in each unit of the banquet idea generation engine as the execution subjects, and the prompt information of each intelligent agent is defined as the input of the execution subject. The content of the prompt information includes role information, input information, task requirements, and required output information. The type and content of the prompt information correspond to each execution subject.

[0037] Specifically, the P111 main control unit embeds a main control agent: determining the corresponding input information as the main control prompt information, including defining the corresponding role information as the main control agent, the input information being the first-level attributes and attribute content of the demand topic, and the task requirement being to determine the brand domain knowledge scope for assisting in banquet design;

[0038] The master control agent is used to understand the task and plan the overall task based on the master control prompts, and outputs the brand domain knowledge scope of the auxiliary banquet design.

[0039] P112 Domain knowledge acquisition unit embeds domain knowledge vector library intelligent agent: determines the corresponding input information as brand domain prompt information, including defining the corresponding role information as brand knowledge agent, the input information is the brand domain knowledge scope of the auxiliary banquet design, and the task requirement is to retrieve and output brand knowledge;

[0040] The domain knowledge vector library agent, based on the brand domain prompt information and combined with the brand domain knowledge vector library, performs in-depth analysis of user characteristics, executes dynamic retrieval, and outputs brand knowledge.

[0041] P113 Creative Generation Unit Embeds Creative Intelligent Agent: Determines the corresponding input information as creative prompt information, including defining the corresponding role information as banquet design creative agent, the input information as brand knowledge and the first-level attributes and attribute content of the demand theme, and the task requirement as generating creative menu plans for the banquet;

[0042] The creative agent outputs creative dish proposals based on creative prompts and a knowledge vector library for the banquet design domain. The creative dish proposals consist of N-level attributes and attribute values ​​of the demand theme.

[0043] The P114 fit evaluation unit embeds a fit evaluation agent: the corresponding input information is determined as the evaluation prompt information, including defining the role information as the aesthetic fit agent, the input information as the first-level attributes and attribute content of the creative dish plan and the demand theme, and the task requirement is to generate the aesthetic fit evaluation result;

[0044] The matching evaluation agent, based on evaluation prompts, introduces reflective reasoning and multi-agent competition mechanisms to generate aesthetic matching evaluation information and provide feedback suggestions. For example, mechanisms such as reflective reasoning and multi-agent competition are introduced to improve the diversity of generated creative ideas and brand matching. The evaluation information includes evaluation results and optimization suggestions, and the types of evaluation results include result confirmation and return for optimization.

[0045] The specific process of the agent-based AI pipeline includes: the output of the master AI agent is the input of the domain knowledge vector library AI agent; the output of the domain knowledge vector library AI agent is the input of the creative AI agent; the output of the creative AI agent is the input of the fit evaluation AI agent; and the evaluation result of the output of the fit evaluation AI agent determines whether to output the banquet creative design project.

[0046] When the evaluation result of the matching evaluation agent is of the type of result confirmation, the creative dish plan is output by the banquet creative generation engine as a banquet creative design project; when the evaluation result is of the type of return optimization, the creative agent is called repeatedly, and after obtaining the creative dish plan output by the creative agent again, the aesthetic matching evaluation information is generated again through the matching evaluation agent.

[0047] The agent-based AI pipeline supports multi-round dialogues and optimizes specific creative solutions for content generation. For example, it can use few-shot learning to improve the quality of creative generation in small-sample scenarios based on past excellent banquet creative cases.

[0048] P120 Brand Domain Knowledge Vector Library: Establishes a data connection with the domain knowledge acquisition unit, and provides domain knowledge related to the user's characteristics when the domain knowledge vector library agent performs a retrieval.

[0049] P130 Banquet Design Domain Knowledge Vector Base: Establishes a data connection with the creative generation unit, and provides banquet design knowledge related to the actual capabilities of the banquet implementer when the creative agent generates creative menu plans;

[0050] The P140 multi-round optimization module is used to define multi-round prompting engineering and feedback mechanisms, and performs self-iteration after the banquet creative generation engine outputs the banquet creative design project.

[0051] In the multi-round optimization module, multiple rounds of prompts and feedback mechanisms (such as brand evaluations and feedback from chefs and experts) are embedded with brand themes and menu ideas to generate and iterate the ideas in multiple rounds, continuously optimizing the creative generation results, meeting the precise needs of high-end brands for banquet ideas, and continuously iterating to improve the quality and efficiency of high-end business banquet creative design.

[0052] P150: Design Output Module: Used to obtain banquet creative design projects output by the banquet creative generation engine and form a banquet creative design scheme.

[0053] The design output module ultimately produces a complete banquet creative execution plan with clear implementation details and creative interpretation. This plan can include multimodal information such as text, images, and structured menus, including the fit between the dishes and the banquet theme, the purpose of ingredient selection, cooking methods, and dining experience, thereby guiding the actual banquet creative team in production and design.

[0054] This invention proposes a banquet creative design system based on a generative large-scale model with multiple agents for banquet creative design, enabling automatic and rapid iteration of creative generation schemes. By utilizing chain of thought and reflection mechanisms, it ensures continuous optimization of creative schemes. A multi-agent collaborative framework is employed, clearly distinguishing the roles of agents in planning, domain knowledge, creative generation, and banquet aesthetic theme fit assessment. Each agent focuses on its specific domain, significantly improving the overall quality and efficiency of creative generation through structured collaboration. Simultaneously, during the banquet scheme creative generation process, the collaboration and decision-making processes among multiple agents can be systematically recorded, making the reasons for banquet creative decisions clear and transparent, facilitating review and optimization. In summary, the proposed agent-based artificial intelligence algorithm can enhance the deep integration of multimodal brand domain knowledge and banquet creative design. Through multi-intelligence planning, multi-round reflection, and collaborative generation mechanisms, it supports an adaptive high-end business banquet creative iteration process with real-time feedback and targeted optimization from the client brand, ultimately achieving the automatic generation of multimodal information for high-end banquet catering art schemes.

[0055] The above-disclosed embodiments are merely a few specific examples of the present invention. However, the present invention is not limited thereto, and any variations that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A banquet creative design system based on generative artificial intelligence, characterized in that, include: Demand Determination Module: Used to interact with user terminals, obtain user requirements, parse the user requirements, and generate a demand theme; the demand theme refers to the service items proposed by the user to the banquet implementer, which consists of multi-level attributes and attribute content; the multi-level attributes include first-level attributes and N-level attributes, the first-level attributes can directly reflect the user's requirements, and the requirements reflected by the first-level attributes are decomposed into dimensions that need to be confirmed during the design, forming N-level attributes; Banquet Creative Generation Engine: Used to identify and run a pipeline of agent-based artificial intelligence agents to generate banquet creative design projects based on the required theme; wherein, the banquet creative generation engine includes a main control unit, a domain knowledge acquisition unit, a creative generation unit, and a fit evaluation unit; each unit embeds multi-level intelligent agents as execution subjects, and defines the prompt information of each intelligent agent as the input of the execution subject; Brand Domain Knowledge Vector Library: Used to provide domain knowledge related to user characteristics; Banquet Design Knowledge Vector Library: Provides banquet design knowledge related to the actual capabilities of banquet implementers; Design Output Module: Used to acquire banquet creative design projects output by the banquet creative generation engine and form banquet creative design schemes.

2. The banquet creative design system according to claim 1, characterized in that, The main control unit is embedded with a main control agent, including: determining the corresponding input information as the main control prompt information, including defining the corresponding role information as the main control agent, the input information being the first-level attributes and attribute content of the demand topic, and the task requirement being to determine the brand domain knowledge scope for assisting in banquet design; The master control agent is used to understand the task and plan the overall task based on the master control prompts, and outputs the brand domain knowledge scope of the auxiliary banquet design.

3. The banquet creative design system according to claim 2, characterized in that, The domain knowledge acquisition unit is embedded with a domain knowledge vector library agent, including: determining the corresponding input information as brand domain prompt information, including defining the corresponding role information as brand knowledge agent, the input information being the brand domain knowledge scope of the auxiliary banquet design, and the task requirement being to retrieve and output brand knowledge; The domain knowledge vector library agent, based on the brand domain prompt information and combined with the brand domain knowledge vector library, performs in-depth analysis of user characteristics, executes dynamic retrieval, and outputs brand knowledge.

4. The banquet creative design system according to claim 3, characterized in that, The creative generation unit embeds a creative intelligent agent, including: determining the corresponding input information as creative prompt information, including defining the corresponding role information as banquet design creative agent, the input information being brand knowledge and the first-level attributes and attribute content, and the task requirement being to generate creative menu plans for the banquet; The creative agent outputs creative dish proposals based on creative prompts and a knowledge vector library for the banquet design domain. The creative dish proposals consist of N-level attributes and attribute values ​​of the demand theme.

5. The banquet creative design system according to claim 4, characterized in that, The matching degree evaluation unit embeds a matching degree evaluation intelligent agent, including: determining the corresponding input information as evaluation prompt information, including defining the role information as aesthetic matching degree Agent, the input information as the first-level attributes and attribute content of the creative dish plan and the demand theme, and the task requirement as generating aesthetic matching degree evaluation results; The matching evaluation agent generates aesthetic matching evaluation information and provides feedback suggestions based on the evaluation prompts. The evaluation information includes evaluation results and optimization suggestions. The evaluation results include result confirmation and return for optimization.

6. The banquet creative design system according to claim 5, characterized in that, When the evaluation result of the matching evaluation agent is of the type of result confirmation, the creative dish plan is output by the banquet creative generation engine as a banquet creative design project; when the evaluation result is of the type of return optimization, the creative agent is called repeatedly, optimization suggestions are input, and the creative dish plan output by the creative agent is obtained again. Then, the aesthetic matching evaluation information is generated again through the matching evaluation agent.

7. The banquet creative design system according to claim 1, characterized in that, The specific process of the agent-based AI pipeline includes: the output of the master AI agent is the input of the domain knowledge vector library AI agent, the output of the domain knowledge vector library AI agent is the input of the creative AI agent, and the output of the creative AI agent is the input of the fit evaluation AI agent; based on the evaluation result of the output of the fit evaluation AI agent, it is determined whether to output the banquet creative design project.

8. The banquet creative design system according to claim 1, characterized in that, The demand determination module determines the content of the first-level attributes and the category of the N-level attributes. In the banquet idea generation engine, the content of the N-level attributes is obtained by each agent AI agent.

Images