A method and system for mass production of creative advertising materials

Abstract

The application relates to the technical field of data processing, and particularly discloses a method and system for generating creative advertisement materials on a large scale, which comprises the following steps: obtaining a candidate set of creative element expressions based on specific element information of multi-dimensional creative element points, and performing rule-based combination on the candidate set of creative element expressions to obtain a plurality of initial creative element combinations; generating a plurality of initial advertisement materials for each initial creative element combination based on a preset AI model; constructing a multi-dimensional creative element network based on the binding depth between the multi-dimensional creative element points; analyzing the evaluation value of each initial advertisement material relative to a multi-dimensional evaluation standard based on the multi-dimensional creative element network; and screening out optimal advertisement materials based on the evaluation value of all initial advertisement materials relative to the multi-dimensional evaluation standard; and combining artificial intelligence to realize large-scale generation of creative advertisement materials.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for large-scale generation of creative advertising materials. Background Technology

[0002] In today's fiercely competitive business environment, advertising, as a crucial means for businesses to promote products and services and attract consumers, directly impacts their market share and economic benefits. With the rapid development of the internet, digital media, and artificial intelligence, advertising channels are becoming increasingly diversified, and consumers are exposed to a massive amount of advertising information daily, making it increasingly difficult for advertisements to stand out. Therefore, creativity plays a pivotal role in advertising production; unique and engaging creative advertising materials can effectively attract consumer attention, enhance brand awareness, and boost product sales. For businesses, especially those with extensive operations and large-scale marketing needs, there is a need to rapidly and massively deploy diverse advertisements across different platforms and targeting different audiences. This necessitates the ability to generate creative advertising materials at scale to meet constantly changing market demands and diversified advertising strategies. Methods and systems for generating creative advertising materials at scale can not only improve advertising production efficiency and reduce labor costs, but also generate rich and diverse targeted advertising materials by mining and integrating various creative elements, helping businesses maintain a competitive edge in the fierce market. With the continuous development of digital marketing, these methods and systems have broad application prospects in the advertising industry, driving advertising production towards a more intelligent and efficient direction.

[0003] However, existing methods for generating advertising creatives lack effective means based on pre-set AI models, making it difficult to generate advertising creatives efficiently and with high quality. Furthermore, due to insufficient understanding and application of the relationships between creative elements, a comprehensive and objective evaluation of advertising creatives is impossible. Ultimately, it becomes difficult to select the optimal advertising creatives that truly meet market demands and corporate goals, impacting advertising effectiveness and corporate marketing efficiency.

[0004] Therefore, this invention proposes a method and system for large-scale generation of creative advertising materials. Summary of the Invention

[0005] This invention provides a method and system for large-scale generation of creative advertising materials. The method generates a candidate set of creative element expressions through multi-dimensional creative element point information and combines them according to rules. This systematically and comprehensively explores creative possibilities, enriches the sources of initial creative element combinations, and lays the foundation for diverse advertising material generation. Initial advertising materials are generated using a pre-set AI model. Leveraging the efficiency and innovation of AI, large-scale production of advertising materials is achieved, saving manpower and time costs. A multi-dimensional creative element network is constructed, revealing the inherent connections between elements based on the binding depth between elements, providing a logical framework for subsequent evaluation. Based on this network, the evaluation values ​​of the initial advertising materials relative to multi-dimensional evaluation standards are analyzed, making the evaluation more scientific and comprehensive, measuring material quality from multiple dimensions. Finally, the optimal advertising materials are selected based on the evaluation values, ensuring that the final output advertising materials are of high quality, better aligned with market demands and marketing goals, and improving advertising effectiveness and commercial value.

[0006] This invention provides a method for large-scale generation of creative advertising materials, comprising:

[0007] Based on the specific element information of multidimensional creative element points, a set of creative element expression candidates is obtained, and the creative element expression candidate set is combined according to rules to obtain multiple initial creative element combinations.

[0008] Multiple initial advertising creatives are generated based on a preset AI model for each initial creative element combination;

[0009] Construct a multidimensional creative element network based on the binding depth between multidimensional creative element points;

[0010] Based on the multidimensional creative element network, the evaluation value of each initial advertising creative is analyzed relative to the multidimensional evaluation criteria;

[0011] The optimal ad creative is selected based on the evaluation values ​​of all initial ad creatives relative to the multidimensional evaluation criteria.

[0012] Optionally, the process of acquiring specific element information for multi-dimensional creative element points includes:

[0013] Based on the user-defined element information of all user-defined creative element points in the multidimensional creative element points, the extended element information of all user-defined creative element points is retrieved. At the same time, the alternative element information of all non-user-defined creative element points is retrieved.

[0014] The user-selected element information of all user-defined creative element points in the extended element information of all user-defined creative element points is deduplicated and summarized with the user-set element information of the corresponding user-defined creative element points to obtain the specific element information of the corresponding user-defined creative element points.

[0015] Set all user-selected element information in the candidate element information of each non-user-defined creative element point to the specific element information of the corresponding non-user-defined creative element point.

[0016] Optionally, a candidate set of creative element expressions is obtained based on the specific element information of the multi-dimensional creative element points, including:

[0017] Obtain detailed information on specific creative elements and conversion strategies from a large number of successful advertising case studies across multiple creative elements.

[0018] The creative element conversion strategies of all successful advertising cases where the specific element information of the corresponding creative element point differs from the specific element information of a single creative element point of the target product / service are taken as all alternative creative element conversion strategies for the corresponding creative element point.

[0019] Eliminate the creative element transformation ideas that have been implemented in similar products and services of the target product and service from all the alternative creative element transformation ideas for each creative element point to obtain all the optional creative element transformation ideas for each creative element point.

[0020] By summarizing all the possible creative element transformation ideas of all creative element points in the multidimensional creative element points, a candidate set of creative element expression is obtained.

[0021] Optionally, the candidate set of creative elements is combined according to rules to obtain multiple initial combinations of creative elements, including:

[0022] Obtain the original constraints of the transformation ideas of each optional creative element in the candidate set of creative element expression, and generate constraint rules based on the original constraints of the transformation ideas of all optional creative elements;

[0023] Based on the constraint rules, the transformation ideas of all optional creative elements for all creative element points in the creative element expression candidate set are arbitrarily combined to obtain multiple initial creative element combinations.

[0024] Optionally, the evaluation value of each initial ad creative relative to the multidimensional evaluation criteria is parsed based on the multidimensional creative element network, including:

[0025] Scoring is performed on a portion of each creative element within the performance range of each initial ad creative, and the evaluation value of each creative element under the multi-dimensional evaluation criteria is obtained for each initial ad creative.

[0026] Based on the multidimensional creative element network, all creative element points whose binding depth with each creative element point is not 0 are identified as all bound element points of the corresponding creative element point;

[0027] Based on the multi-level binding depth threshold, all binding element points of each creative element point are divided to obtain all binding element points under each binding depth level of each creative element point;

[0028] The quotient of the binding depth of each bound element point of each creative element point to the corresponding creative element point and the sum of the binding depths of all bound element points to the corresponding creative element point at the same binding depth level is used as the weight of the corresponding bound element point at the corresponding binding depth level.

[0029] The sum of the products of each evaluation value of all bound element points at each binding depth level in each initial ad creative and the weight at the corresponding binding depth level, and the weight at the corresponding binding depth level, is taken as the additional effect of all bound element points at each binding depth level on the corresponding evaluation standard of the corresponding creative element point.

[0030] The sum of the evaluation value of each creative element point under the multidimensional evaluation criteria in each initial ad creative and the additional effect of all bound element points under all binding depth levels on the corresponding evaluation criteria of the corresponding creative element point is taken as the evaluation value of each creative element point under the multidimensional evaluation criteria in each initial ad creative.

[0031] The evaluation value of each initial ad creative is determined based on the evaluation value of each creative element under the multidimensional evaluation criteria in each initial ad creative.

[0032] Optionally, the evaluation value of each initial ad creative is determined relative to the multidimensional evaluation criteria based on the evaluation value of each creative element point under the multidimensional evaluation criteria in each initial ad creative, including:

[0033] Based on the average degree of the total number of interval elements between each creative element point and all edge creative element points in the multidimensional creative element network, as well as the degree centrality, betweenness centrality, and proximity centrality of the corresponding creative element point in the multidimensional creative element network, the core position value of each creative element point in the multidimensional creative element network is calculated.

[0034] Based on the evaluation value of each creative element under the multidimensional evaluation criteria in each initial ad creative and its core position value in the multidimensional creative element network, the effectiveness weight of each creative element in the multidimensional creative element network is calculated.

[0035] Based on the effectiveness weights of all creative elements in the multidimensional creative element network and the evaluation values ​​of all creative elements under the multidimensional evaluation criteria in each initial ad creative, the evaluation value of each initial ad creative relative to the multidimensional evaluation criteria is calculated.

[0036] Optionally, based on the average degree of the total number of margin points between each creative element point and all edge creative element points in the multidimensional creative element network, and the degree centrality, betweenness centrality, and proximity centrality of the corresponding creative element point in the multidimensional creative element network, the core position value of each creative element point in the multidimensional creative element network is calculated, including:

[0037] When there is a unique creative element point in the multidimensional creative element network that has the largest average degree of the total number of interval element points between all edge creative element points, then the bounding lines of creative element points at all levels of the corresponding creative element point are determined in the multidimensional creative element network, and the core position value of each creative element point in the multidimensional creative element network is determined based on the bounding lines of creative element points at all levels.

[0038] When there is more than one creative element point in the multidimensional creative element network that has the highest average degree of the total number of interval element points among all edge creative element points, the bounding line of creative element points at all levels for each creative element point with the highest average degree is determined.

[0039] The centrality value of each creative element point in the multidimensional creative element network is calculated based on its degree centrality, betweenness centrality, and proximity centrality.

[0040] Based on the binding depth between each creative element point with the maximum average degree and all creative element points in the enclosing line of creative element points at all levels, and the product of the centrality values ​​of all creative element points in the enclosing line of creative element points at all levels in the multidimensional creative element network, a comprehensive matrix of the corresponding creative element points is generated.

[0041] The core position value of each creative element in the multidimensional creative element network is calculated based on the comprehensive matrix of each creative element point with the maximum average degree.

[0042] Optionally, the core position value of each creative element point in the multidimensional creative element network is calculated based on the comprehensive matrix of each creative element point with the maximum average degree, including:

[0043] Determine the comprehensive similarity between the eigenvector corresponding to the largest eigenvalue of the comprehensive matrix of each creative element point with the maximum average degree and the row vector corresponding to the enclosing line of creative element points at all levels in the comprehensive matrix;

[0044] Based on the creative element points with the highest average degree, the enclosing line of all levels of creative element points with the highest comprehensive similarity determines the core position value of each creative element point in the multidimensional creative element network.

[0045] Optionally, based on the evaluation values ​​of all initial ad creatives relative to the multidimensional evaluation criteria, the optimal ad creatives are selected, including:

[0046] Based on the evaluation value of each initial ad creative relative to the multidimensional evaluation criteria, a comprehensive evaluation value for each initial ad creative is determined;

[0047] The initial ad creative with the highest overall score among all initial ad creatives is considered the optimal ad creative.

[0048] This invention provides a system for large-scale generation of creative advertising materials, comprising:

[0049] The element information combination module is used to obtain a set of creative element expression candidates based on the specific element information of multi-dimensional creative element points, and to combine the creative element expression candidate sets according to rules to obtain multiple initial creative element combinations.

[0050] The ad creative generation module is used to generate multiple initial ad creatives based on a preset AI model for each initial creative element combination;

[0051] The element network construction module is used to construct a multi-dimensional creative element network based on the binding depth between multi-dimensional creative element points;

[0052] The advertising effectiveness evaluation module is used to analyze the evaluation value of each initial advertising creative relative to the multidimensional evaluation criteria based on the multidimensional creative element network.

[0053] The ad creative selection module is used to select the best ad creatives based on the evaluation values ​​of all initial ad creatives relative to the multidimensional evaluation criteria.

[0054] The beneficial effects of this invention compared to existing technologies are as follows: By generating a candidate set of creative element expressions through multi-dimensional creative element point information and combining them according to rules, it can systematically and comprehensively explore creative possibilities, enrich the sources of initial creative element combinations, and lay the foundation for the generation of diversified advertising materials. It utilizes a preset AI model to generate initial advertising materials, leveraging the efficiency and innovation of AI to achieve large-scale production of advertising materials, saving manpower and time costs. It constructs a multi-dimensional creative element network, revealing the intrinsic connections between elements based on the binding depth between elements, providing a logical framework for subsequent evaluation. Based on this network, it analyzes the evaluation values ​​of the initial advertising materials relative to multi-dimensional evaluation standards, making the evaluation more scientific and comprehensive, measuring material quality from multiple dimensions. Finally, it selects the optimal advertising materials based on the evaluation values, ensuring that the final output advertising materials are of high quality, better aligned with market demands and marketing goals, and improving advertising effectiveness and commercial value.

[0055] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained through the structures particularly pointed out in this application.

[0056] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0057] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0058] Figure 1 This is a flowchart illustrating the method and process for large-scale generation of creative advertising materials in an embodiment of the present invention.

[0059] Figure 2 This is a flowchart illustrating the process of acquiring specific element information of multidimensional creative element points in an embodiment of the present invention.

[0060] Figure 3 This is a flowchart illustrating the process of obtaining a candidate set of creative element expressions in an embodiment of the present invention. Detailed Implementation

[0061] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0062] This invention provides an implementation method for large-scale generation of creative advertising materials, comprising:

[0063] Based on the specific element information of multidimensional creative element points, a set of creative element expression candidates is obtained, and the creative element expression candidate set is combined according to rules to obtain multiple initial creative element combinations.

[0064] Multiple initial advertising creatives are generated based on a preset AI model for each initial creative element combination;

[0065] Construct a multidimensional creative element network based on the binding depth between multidimensional creative element points;

[0066] Based on the multidimensional creative element network, the evaluation value of each initial advertising creative is analyzed relative to the multidimensional evaluation criteria;

[0067] The optimal ad creative is selected based on the evaluation values ​​of all initial ad creatives relative to the multidimensional evaluation criteria.

[0068] In this embodiment, multi-dimensional creative elements refer to the basic units that describe advertising creativity from multiple dimensions. Each dimension contains specific creative-related elements, such as the selling points of the product or service, the audience's emotional starting point, the landing platform, the advertising format, and the technical implementation.

[0069] In this embodiment, the specific element information of the multi-dimensional creative element point is the specific performance information under the multi-dimensional creative element point: for example, the product's selling point is "quick-drying + hair care", the audience's emotional starting point is "anxiety relief", the landing platform is "Douyin", the advertising format is "short video", and the technical implementation is "AI-generated content", etc.

[0070] In this embodiment, a preset AI model, such as the existing D&S-AI Video or OpenAI GPT-3 model, is provided, which has the ability to generate advertising materials based on the combination of initial creative elements input, and can transform the combination of initial creative elements input into specific advertising materials.

[0071] In this embodiment, multiple initial advertising materials are generated based on a preset AI model for each initial creative element combination: each initial creative element combination is input into the preset AI model, and the model outputs multiple corresponding initial advertising materials based on its own algorithm and training results. For example, inputting a combination of creative elements including the selling points of a specific product or service, the audience's emotional starting point, the landing platform, the advertising format, and the technical implementation, the model generates different advertising materials.

[0072] In this embodiment, the binding depth between multidimensional creative elements is a numerical value that measures the degree of correlation between multidimensional creative elements. The higher the depth, the closer the connection between the two creative elements. For example, if the creative elements of the landing platform and the advertising format always appear together and cooperate closely in many successful cases, the binding depth between them is high.

[0073] In this embodiment, a multi-dimensional creative element network is constructed based on the binding depth between multi-dimensional creative element points: the network structure is constructed with multi-dimensional creative element points as nodes and the binding depth between them as the edge weights. For example, if there are multiple creative element points such as product / service selling points, audience emotional starting points, landing platforms, advertising formats, and technical implementations, and the binding depth between advertising formats and landing platforms is 0.8, the binding depth between landing platforms and audience emotional starting points is 0.6, and the binding depth between selling points and audience emotional starting points is 0.5, then a network reflecting their relationships can be constructed accordingly.

[0074] In this embodiment, the multidimensional evaluation criteria are standards for evaluating initial advertising creatives from multiple dimensions, which may include dimensions such as attractiveness, relevance, innovativeness, communicability, and memorability, to comprehensively measure the quality of advertising creatives.

[0075] In this embodiment, the evaluation value of each initial advertising creative relative to the multidimensional evaluation criteria is a quantitative value obtained through a series of calculations, reflecting the performance of the initial advertising creative under the multidimensional evaluation criteria.

[0076] Furthermore, to enhance the alignment between the entire creative material generation process and user needs, and to optimize the acquisition of creative element information, the process of acquiring specific element information for multi-dimensional creative element points is referenced. Figure 2 ,include:

[0077] Based on the user-defined element information of all user-defined creative element points in the multidimensional creative element points, the extended element information of all user-defined creative element points is retrieved. At the same time, the alternative element information of all non-user-defined creative element points is retrieved.

[0078] The user-selected element information of all user-defined creative element points in the extended element information of all user-defined creative element points is deduplicated and summarized with the user-set element information of the corresponding user-defined creative element points to obtain the specific element information of the corresponding user-defined creative element points.

[0079] Set all user-selected element information in the candidate element information of each non-user-defined creative element point to the specific element information of the corresponding non-user-defined creative element point.

[0080] In this embodiment, user-defined creative elements are those determined by the user based on their own needs and ideas for advertising creativity. For example, if a user wants the advertisement to highlight a unique feature of the product, the relevant creative elements set around that feature are user-defined creative elements, reflecting the user's specific direction for the advertising creativity.

[0081] In this embodiment, user-defined element information refers to the specific information provided by the user for a given creative element. For example, if the user defines "product advantages" as a creative element, then the user-defined element information might be a description of the product's specific advantages, such as "the product has a fast charging function, which can charge 80% of the battery in half an hour," providing a foundation for subsequent retrieval and expansion of element information.

[0082] In this embodiment, based on the user-defined element information of all user-defined creative element points in the multi-dimensional creative element points, the system retrieves the extended element information of all user-defined creative element points: according to the setting element information provided by the user for each user-defined creative element point, the system searches in relevant databases or information sources to obtain more related extended information. For example, if the user-defined element information indicates that the product's advantage is fast charging, the system will retrieve more promotional angles and expressions of similar products regarding fast charging as extended element information, such as descriptions of charging speed comparisons and charging technology principles.

[0083] In this embodiment, alternative element information for non-user-defined creative elements: For creative elements that are not actively set by the user, the system provides pre-prepared alternative element information. For example, in advertising creatives, if the user has not set the creative element regarding the overall style, the system will provide multiple alternative element information such as minimalist style, retro style, and tech style to enrich the possibilities of advertising creatives.

[0084] In this embodiment, user-selected element information refers to the portion of information that the user chooses and actually uses, whether it's from the extended element information of user-defined creative element points or from the alternative element information of non-user-defined creative element points. For example, if a user selects "tech-savvy style" from the alternative element information of various advertising styles, this is user-selected element information, and it will ultimately participate in the construction of the advertising creative.

[0085] Furthermore, in order to build a rich and unique creative resource library and provide ample and high-quality creative reserves for generating eye-catching and outstanding creative advertising materials, a candidate set of creative element expressions is obtained based on the specific element information of multi-dimensional creative element points, referencing... Figure 3 ,include:

[0086] Obtain detailed information on specific creative elements and conversion strategies from a large number of successful advertising case studies across multiple creative elements.

[0087] The creative element conversion strategies of all successful advertising cases where the specific element information of the corresponding creative element point differs from the specific element information of a single creative element point of the target product / service are taken as all alternative creative element conversion strategies for the corresponding creative element point.

[0088] Eliminate the creative element transformation ideas that have been implemented in similar products and services of the target product and service from all the alternative creative element transformation ideas for each creative element point to obtain all the optional creative element transformation ideas for each creative element point.

[0089] By summarizing all the possible creative element transformation ideas of all creative element points in the multidimensional creative element points, a candidate set of creative element expression is obtained.

[0090] In this embodiment, a successful advertising case refers to an advertising example that has achieved good results in the market and achieved goals such as attracting consumer attention, increasing product sales, and enhancing brand awareness.

[0091] In this embodiment, we acquire specific element information and creative element conversion strategies from a large number of successful advertising cases under multi-dimensional creative element points: we collect detailed information on numerous successful advertisements across multiple creative element dimensions, and explore thought processes on how to apply these creative elements to advertising and achieve effective dissemination. For example, analyzing a successful advertisement, its specific element information under multi-dimensional creative element points might include the theme of "youthful vitality," the style of "trendy fashion," and the copywriting of "Enjoy youth, set off now," etc. The creative element conversion strategy is how to integrate these elements around the preferences of young people and disseminate them through specific media channels.

[0092] In this embodiment, the creative element transformation approach is the thinking direction and method of transforming creative elements from concepts into actual advertising expressions. For example, for the creative element of "green and environmentally friendly", the transformation approach might be to convey the concept of environmental protection to consumers through vivid visuals and concise copy by showcasing the product's environmentally friendly production process and the use of environmentally friendly materials, thereby attracting environmentally conscious consumers.

[0093] In this embodiment, the target product / service refers to the specific product or service that this advertising campaign is aimed at. For example, for a newly launched smartphone, the generated advertising materials are all centered around promoting this smartphone.

[0094] In this embodiment, the creative element conversion approach already implemented in similar products and services to the target product / service refers to creative conversion methods that have already been implemented and applied in advertisements for other products or services of the same type as the target product or service. For example, if the target product is a new running shoe from a certain brand, and other brands of similar products have already used the creative element conversion approach of "celebrity endorsement showcasing the running shoe's performance" in their advertising, this approach, already adopted by similar products, may be removed when generating advertising materials for the target product to pursue creative uniqueness. This increases the innovation of the generated advertising materials and avoids homogeneity with similar advertisements.

[0095] Furthermore, to balance the feasibility of the creative ideas with their diversity, and to lay a solid foundation for generating high-quality, creative advertising materials, the candidate set of creative elements was combined according to regulations to obtain several initial combinations of creative elements, including:

[0096] Obtain the original constraints for the transformation ideas of each optional creative element in the candidate set of creative element expressions, and generate constraint rules based on the original constraints for the transformation ideas of all optional creative elements:

[0097] Each potential creative element conversion strategy has its own initial constraints, which may stem from actual business needs, technological limitations, market positioning, and other factors. For example, a potential creative element conversion strategy for "high-end skincare product advertising" might have initial constraints such as a limited budget, a target audience of high-income women, and a focus on natural ingredients. After collecting the initial constraints of all these strategies in the creative element expression candidate set, they are summarized, analyzed, and integrated to generate a set of general constraint rules. For example, from the budget constraints of multiple strategies, an overall budget allocation rule can be summarized; from the audience positioning condition, rules for the creative style and promotional focus targeting high-income women can be determined. These constraint rules ensure that the subsequently generated creative combinations are realistically feasible and commercially viable.

[0098] Based on the constraint rules, all possible creative element transformation ideas of all creative element points in the creative element expression candidate set are arbitrarily combined to obtain multiple initial creative element combinations:

[0099] Based on the previously generated constraints, the conversion strategies for all available creative elements corresponding to each creative element point in the creative element expression candidate set are combined. During the combination process, any combination is allowed as long as it conforms to the constraints. For example, suppose the theme creative element points have two options: "Natural Beauty" and "Technological Rejuvenation"; the style creative element points have two options: "Simple and Fresh" and "Luxury and Grand"; and the copywriting creative element points have two options: "Luxury Skincare" and "Embark on a Technological Skincare Journey." The constraints allow for free combination while ensuring the overall budget and target audience. Then, multiple initial creative element combinations can be created, such as "Natural Beauty - Simple and Fresh - Luxurious Skincare" and "Technological Rejuvenation - Luxurious and Grand - Embark on a Technological Skincare Journey," providing a diverse creative foundation for subsequent generation of advertising materials using a preset AI model.

[0100] Furthermore, to make the evaluation of initial advertising creatives more comprehensive and accurate, a multidimensional creative element network was used to analyze the evaluation value of each initial advertising creative relative to the multidimensional evaluation criteria, including:

[0101] Scoring is performed on a portion of each creative element within the performance range of each initial ad creative, and the evaluation value of each creative element under the multi-dimensional evaluation criteria is obtained for each initial ad creative.

[0102] Based on the multidimensional creative element network, all creative element points whose binding depth with each creative element point is not 0 are identified as all bound element points of the corresponding creative element point;

[0103] Based on the multi-level binding depth threshold, all binding element points of each creative element point are divided to obtain all binding element points under each binding depth level of each creative element point;

[0104] The quotient of the binding depth of each bound element point of each creative element point to the corresponding creative element point and the sum of the binding depths of all bound element points to the corresponding creative element point at the same binding depth level is used as the weight of the corresponding bound element point at the corresponding binding depth level.

[0105] The sum of the products of each evaluation value of all bound element points at each binding depth level in each initial ad creative and the weight at the corresponding binding depth level, and the weight at the corresponding binding depth level, is taken as the additional effect of all bound element points at each binding depth level on the corresponding evaluation standard of the corresponding creative element point.

[0106] The sum of the evaluation value of each creative element point under the multidimensional evaluation criteria in each initial ad creative and the additional effect of all bound element points under all binding depth levels on the corresponding evaluation criteria of the corresponding creative element point is taken as the evaluation value of each creative element point under the multidimensional evaluation criteria in each initial ad creative.

[0107] The evaluation value of each initial ad creative is determined based on the evaluation value of each creative element under the multidimensional evaluation criteria in each initial ad creative.

[0108] In this embodiment, a subset of advertising materials within the performance range of each creative element point in each initial advertising creative is scored to obtain the evaluation value of each creative element point under multi-dimensional evaluation criteria in each initial advertising creative: For each initial advertising creative, a subset of sample advertising materials within the performance range involved in each creative element point are scored according to a preset model based on multi-dimensional evaluation criteria (such as attractiveness, relevance, and originality). For example, for an initial advertising creative with the theme of "fashion clothing," within the performance range of the creative element point "style," a subset of visual clips showcasing clothing styles are selected. The model scores these clips based on the attractiveness dimension criteria, considering aspects such as color matching and design uniqueness, ultimately obtaining the evaluation value of each creative element point under the corresponding multi-dimensional evaluation criteria in this initial advertising creative. This approach allows for a preliminary evaluation of the quality of advertising creatives from various creative element point dimensions. The preset model is trained in advance using a large number of advertising creatives scored by experts from the perspective of multi-dimensional evaluation criteria.

[0109] In this embodiment, the multi-level binding depth threshold is a series of pre-set values ​​used to define the level boundaries of binding depth between creative element points. For example, the binding depth thresholds are set to 0.2, 0.5, and 0.8. When the binding depth between creative element point A and creative element point B is between 0 and 0.2, it belongs to the first level; between 0.2 and 0.5, it belongs to the second level; between 0.5 and 0.8, it belongs to the third level; and greater than 0.8, it belongs to the fourth level.

[0110] In this embodiment, all bound element points of each creative element point are divided based on a multi-level binding depth threshold to obtain all bound element points at each binding depth level for each creative element point:

[0111] Based on the previously established multi-level binding depth thresholds, all binding element points with a binding depth not equal to 0 for each creative element point are categorized. Assume the creative element point "Brand Positioning" has binding element points "Target Audience," "Product Features," and "Promotion Channels." Its binding depth with "Target Audience" is 0.6, with "Product Features" it is 0.3, and with "Promotion Channels" it is 0.7. According to the multi-level binding depth thresholds of 0.2, 0.5, and 0.8, "Target Audience" and "Promotion Channels" belong to a higher binding depth level, while "Product Features" belongs to a medium binding depth level.

[0112] In this embodiment, the weight of the binding depth level is a numerical value assigned to each binding depth level to represent its relative importance in the evaluation process. For example, the binding depth levels are divided into three levels: high, medium, and low. The high level is assigned a weight of 0.6, the medium level a weight of 0.3, and the low level a weight of 0.1. When calculating the initial evaluation value of the advertising creative, the binding depth level weight is used to adjust the contribution of different levels of binding elements to the evaluation of the creative elements. The higher the weight, the greater the impact of that level of binding element on the final evaluation value, thus more accurately considering the role of the relationship between various elements in the evaluation of the advertising creative.

[0113] Furthermore, to comprehensively and objectively evaluate creative materials and aid in the selection of high-quality ad creatives, the evaluation value of each initial ad creative relative to the multidimensional evaluation criteria is determined based on the evaluation value of each creative element under the multidimensional evaluation criteria in each initial ad creative, including:

[0114] Based on the average degree of the total number of interval elements between each creative element point and all edge creative element points in the multidimensional creative element network, as well as the degree centrality, betweenness centrality, and proximity centrality of the corresponding creative element point in the multidimensional creative element network, the core position value of each creative element point in the multidimensional creative element network is calculated.

[0115] Based on the evaluation value of each creative element under the multidimensional evaluation criteria in each initial ad creative and its core position value in the multidimensional creative element network, the effectiveness weight of each creative element in the multidimensional creative element network is calculated.

[0116] Based on the effectiveness weights of all creative element points in the multidimensional creative element network, the evaluation values ​​of all creative element points under the multidimensional evaluation criteria in each initial ad creative are weighted and summed to obtain the evaluation value of each initial ad creative relative to the multidimensional evaluation criteria.

[0117] In this embodiment, the average degree of spacing between each creative element point and all edge creative element points in the multidimensional creative element network is:

[0118] In a multidimensional creative element network, edge creative element points can be understood as creative element points located at the relative edges of the network. For a specific creative element point, the total number of element points between it and all edge creative element points is calculated, and then the average of these totals is taken. For example, suppose there is a simple multidimensional creative element network containing creative element points A, B, C, D, and E, where D and E are edge creative element points. Creative element point A is 2 element points away from D (as with B and C), and 1 element point away from E (as with B). Then the total number of element points between A and all edge creative element points is 3, and its average is 3 - 3 ÷ 2 = 1.5. The average of the absolute value of the difference between the total number of element points between all edge points and this average, and the quotient of the average, is taken as the average degree. This average degree reflects the centrality of the creative element point's position in the network.

[0119] In this embodiment, the effectiveness weight of each creative element in the multidimensional creative element network is calculated based on its evaluation value under the multidimensional evaluation criteria in each initial ad creative and its core position value in the multidimensional creative element network. For example, assuming a creative element in an initial ad creative has an average evaluation value of 8 points (out of 10) under the multidimensional evaluation criteria, its core position value is calculated to be 0.6 (hypothetically). The effectiveness weight can be calculated using the formula: Effectiveness Weight = Evaluation Value × Core Position Value ÷ (Sum of the products of all creative element evaluation values ​​and core position values). For example, if there are two other creative element points with evaluation values ​​of 6 and 7 points respectively, and core position values ​​of 0.4 and 0.5 respectively, then the effectiveness weight of the first creative element point = 8 × 0.6 ÷ (8 × 0.6 + 6 × 0.4 + 7 × 0.5) ≈ 0.37. The effectiveness weight calculated in this way comprehensively considers the performance of creative element points in the ad creative and their importance in the network structure, and is used to more accurately calculate the evaluation value of the initial ad creative relative to the multidimensional evaluation criteria.

[0120] Furthermore, in order to comprehensively and meticulously measure the coreness of each creative element in the network, based on the average degree of the number of margin elements between each creative element and all marginal creative element points in the multidimensional creative element network, and the degree centrality, betweenness centrality, and proximity centrality of the corresponding creative element in the multidimensional creative element network, the core position value of each creative element in the multidimensional creative element network is calculated, including:

[0121] When there is a unique creative element point in the multidimensional creative element network that has the largest average degree of the total number of interval element points between all edge creative element points, then the bounding lines of creative element points at all levels of the corresponding creative element point are determined in the multidimensional creative element network, and the core position value of each creative element point in the multidimensional creative element network is determined based on the bounding lines of creative element points at all levels.

[0122] When there is more than one creative element point in the multidimensional creative element network that has the highest average degree of the total number of interval element points among all edge creative element points, the bounding line of creative element points at all levels for each creative element point with the highest average degree is determined.

[0123] The average of the degree centrality, betweenness centrality, and proximity centrality of each creative element point in the multidimensional creative element network is taken as the centrality value of each creative element point in the multidimensional creative element network.

[0124] Based on the binding depth between each creative element point with the maximum average degree and all creative element points in the enclosing line of creative element points at all levels, and the product of the centrality values ​​of all creative element points in the enclosing line of creative element points at all levels in the multidimensional creative element network, a comprehensive matrix of the corresponding creative element points is generated.

[0125] The core position value of each creative element in the multidimensional creative element network is calculated based on the comprehensive matrix of each creative element point with the maximum average degree.

[0126] In this embodiment, the bounding lines of creative element points at all levels corresponding to the creative element points are determined in the multidimensional creative element network:

[0127] In a multidimensional creative element network, a specific creative element point is taken as the core. Based on the degree of connection (i.e., binding depth) with this core element point, different levels of creative element points are divided from near to far. These different levels of creative element points constitute the bounding lines of all levels of creative element points for the corresponding creative element point. For example, creative element point A has a high binding depth with B and C, a secondary binding depth with D and E, and a low binding depth with F and G. Then, B and C can be considered as the first layer of bounding line elements, D and E as the second layer, and F and G as the third layer. This set of element points at different levels forms the bounding line of creative element point A. It helps to visually demonstrate the distribution of other element points related to the core creative element point.

[0128] Alternatively, the boundaries of creative element points can be divided into all levels based on the total number of edges between them. For example, the line connecting all creative element points with a total of 1 edge to the creative element point is called the first layer of creative element point boundaries; the line connecting all creative element points with a total of 2 edge to the creative element point is called the second layer of creative element point boundaries; the line connecting all creative element points with a total of 3 edge to the creative element point is called the third layer of creative element point boundaries, and so on.

[0129] In this embodiment, the core position value of each creative element point in the multidimensional creative element network is determined based on the enclosing lines of creative element points at all levels. The core position value is calculated by comprehensively considering the relationship between element points at each level and the core creative element point, as well as the overall network structure, according to the determined enclosing lines. For example, levels closer to the core creative element point (i.e., with higher binding depth) are assigned higher weights. Then, considering factors such as the number of element points at each level and the importance of these element points in the network (e.g., degree centrality, betweenness centrality, proximity centrality, etc.), the core position value is calculated using a specific formula. For instance, element points B and C on the first enclosing line of core creative element point A have high degree centrality and a binding depth of 0.8 with the core creative element point, assigned a weight of 0.5; element points D and E on the second enclosing line have moderate degree centrality and a binding depth of 0.5 with the core creative element point, assigned a weight of 0.3; and element points F and G on the third enclosing line have low degree centrality and a binding depth of 0.3 with the core creative element point, assigned a weight of 0.2. The core position value of creative element point A is obtained by calculating the sum of the binding depth and weight products between each layer of element points and the core creative element point. This core position value reflects the relative importance of the creative element point in the entire multidimensional creative element network.

[0130] In this embodiment, a comprehensive matrix of the corresponding creative element points is generated based on the binding depth between each creative element point with the maximum average degree and all creative element points in the enclosing line of creative element points at all levels, and the product of the centrality values ​​of all creative element points in the enclosing line of creative element points at all levels in the multidimensional creative element network.

[0131] Using these products as elements, construct a matrix, which is the comprehensive matrix corresponding to the creative element points. For example, the creative element point X with the highest average intensity has element points a and b in its first layer of enclosing lines, with binding depths of 0.7 and 0.6 respectively, and centrality values ​​of 0.4 and 0.5 respectively; the second layer of enclosing lines has element points c and d, with binding depths of 0.5 and 0.4 respectively, and centrality values ​​of 0.3 and 0.35 respectively. Then the elements in the first row of the comprehensive matrix are 0.7×0.4 and 0.6×0.5, the elements in the second row are 0.5×0.3 and 0.4×0.35, and so on.

[0132] Furthermore, in order to accurately measure the core position of each creative element in the multidimensional creative element network from the perspective of matrix characteristics and inter-element relationships, the core position value of each creative element in the multidimensional creative element network is calculated based on the comprehensive matrix of each creative element with the maximum average degree, including:

[0133] The comprehensive similarity between the eigenvector corresponding to the maximum eigenvalue of the comprehensive matrix of each creative element point with the maximum average degree and the row vectors corresponding to the enclosing lines of creative element points at all levels in the comprehensive matrix is ​​determined, that is, the mean of the cosine similarity between the eigenvector corresponding to the maximum eigenvalue and these row vectors.

[0134] Based on the creative element points with the highest average degree, the enclosing line of all levels of creative element points with the highest comprehensive similarity determines the core position value of each creative element point in the multidimensional creative element network.

[0135] Furthermore, to achieve efficient and accurate creative selection, the optimal creative is selected based on the evaluation values ​​of all initial ad creatives relative to a multi-dimensional evaluation criterion, including:

[0136] Based on the evaluation value of each initial ad creative relative to the multidimensional evaluation criteria, a comprehensive evaluation value for each initial ad creative is determined;

[0137] The initial ad creative with the highest overall score among all initial ad creatives is considered the optimal ad creative.

[0138] In this embodiment, the comprehensive evaluation value of each initial advertising creative is determined based on its evaluation value relative to the multi-dimensional evaluation criteria. This means comprehensively considering the evaluation values ​​of each initial advertising creative under multiple evaluation criteria (e.g., weighted summation) to obtain a single numerical value that can reflect the overall quality or effectiveness of the advertising creative. For example, the multi-dimensional evaluation criteria may include four dimensions: attractiveness, relevance, innovation, and emotional resonance. An initial advertising creative has an evaluation value of 8 points in the attractiveness dimension, 7 points in the relevance dimension, 6 points in the innovation dimension, and 7 points in the emotional resonance dimension. Assuming that, based on experience or a specific algorithm, these four dimensions are assigned weights as follows: attractiveness 0.3, relevance 0.3, innovation 0.2, and emotional resonance 0.2, then the comprehensive evaluation value of this initial advertising creative is calculated as: (8×0.3+7×0.3+6×0.2+7×0.2)=7.1 points. In this way, the evaluation values ​​under the multi-dimensional evaluation criteria are integrated into a single comprehensive evaluation value, facilitating comparison and selection of all initial advertising creatives to determine the optimal advertising creative.

[0139] In addition, the present invention also provides an implementation method for a system for large-scale generation of creative advertising materials, comprising:

[0140] The element information combination module is used to obtain a set of creative element expression candidates based on the specific element information of multi-dimensional creative element points, and to combine the creative element expression candidate sets according to rules to obtain multiple initial creative element combinations.

[0141] The ad creative generation module is used to generate multiple initial ad creatives based on a preset AI model for each initial creative element combination;

[0142] The element network construction module is used to construct a multi-dimensional creative element network based on the binding depth between multi-dimensional creative element points;

[0143] The advertising effectiveness evaluation module is used to analyze the evaluation value of each initial advertising creative relative to the multidimensional evaluation criteria based on the multidimensional creative element network.

[0144] The ad creative selection module is used to select the best ad creatives based on the evaluation values ​​of all initial ad creatives relative to the multidimensional evaluation criteria.

[0145] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for large-scale generation of creative advertising materials, characterized in that, include: Based on the specific element information of multidimensional creative element points, a set of creative element expression candidates is obtained, and the creative element expression candidate set is combined according to rules to obtain multiple initial creative element combinations. Multiple initial advertising creatives are generated based on a preset AI model for each initial creative element combination; Construct a multidimensional creative element network based on the binding depth between multidimensional creative element points; Based on the multidimensional creative element network, the evaluation value of each initial advertising creative is analyzed relative to the multidimensional evaluation criteria; The optimal ad creative is selected based on the evaluation values ​​of all initial ad creatives relative to the multidimensional evaluation criteria. Among them, the evaluation value of each initial advertising creative relative to the multidimensional evaluation criteria is obtained based on the multidimensional creative element network, including: Scoring is performed on a portion of each creative element within the performance range of each initial ad creative to obtain an initial evaluation value for each creative element under a multi-dimensional evaluation standard in each initial ad creative. Based on the multidimensional creative element network, all creative element points whose binding depth with each creative element point is not 0 are identified as all bound element points of the corresponding creative element point; Based on the multi-level binding depth threshold, all binding element points of each creative element point are divided to obtain all binding element points under each binding depth level of each creative element point; The quotient of the binding depth of each bound element point of each creative element point to the corresponding creative element point and the sum of the binding depths of all bound element points to the corresponding creative element point at the same binding depth level is used as the weight of the corresponding bound element point at the corresponding binding depth level. The sum of the products of each evaluation value of all bound element points at each binding depth level in each initial ad creative and the weight at the corresponding binding depth level, and the weight at the corresponding binding depth level, is taken as the additional effect of all bound element points at each binding depth level on the corresponding evaluation standard of the corresponding creative element point. The sum of the initial evaluation value of each creative element point under the multidimensional evaluation criteria in each initial ad creative and the additional effectiveness of all bound element points under all binding depth levels for the corresponding evaluation criteria of the corresponding creative element point is taken as the evaluation value of each creative element point under the multidimensional evaluation criteria in each initial ad creative. The evaluation value of each initial ad creative is determined based on the evaluation value of each creative element under the multidimensional evaluation criteria in each initial ad creative.

2. The method for large-scale generation of creative advertising materials according to claim 1, characterized in that, The process of acquiring specific element information for multidimensional creative elements includes: Based on the user-defined element information of all user-defined creative element points in the multidimensional creative element points, the extended element information of all user-defined creative element points is retrieved. At the same time, the alternative element information of all non-user-defined creative element points is retrieved. The user-selected element information of all user-defined creative element points in the extended element information of all user-defined creative element points is deduplicated and summarized with the user-set element information of the corresponding user-defined creative element points to obtain the specific element information of the corresponding user-defined creative element points. Set all user-selected element information in the candidate element information of each non-user-defined creative element point to the specific element information of the corresponding non-user-defined creative element point.

3. The method for large-scale generation of creative advertising materials according to claim 1, characterized in that, Based on the specific element information of multi-dimensional creative element points, a candidate set of creative element expressions is obtained, including: Obtain detailed information on specific creative elements and conversion strategies from a large number of successful advertising case studies across multiple creative elements. The creative element conversion strategies of all successful advertising cases where the specific element information of the corresponding creative element point differs from the specific element information of a single creative element point of the target product / service are taken as all alternative creative element conversion strategies for the corresponding creative element point. Eliminate the creative element transformation ideas that have been implemented in similar products and services of the target product and service from all the alternative creative element transformation ideas for each creative element point to obtain all the optional creative element transformation ideas for each creative element point. By summarizing all the possible creative element transformation ideas of all creative element points in the multidimensional creative element points, a candidate set of creative element expression is obtained.

4. The method for large-scale generation of creative advertising materials according to claim 3, characterized in that, The candidate set of creative elements is combined according to rules to obtain multiple initial combinations of creative elements, including: Obtain the original constraints of the transformation ideas of each optional creative element in the candidate set of creative element expression, and generate constraint rules based on the original constraints of the transformation ideas of all optional creative elements; Based on the constraint rules, the transformation ideas of all optional creative elements for all creative element points in the creative element expression candidate set are arbitrarily combined to obtain multiple initial creative element combinations.

5. The method for large-scale generation of creative advertising materials according to claim 1, characterized in that, The evaluation value of each initial ad creative is determined based on the evaluation value of each creative element under the multidimensional evaluation criteria in each initial ad creative, including: Based on the average degree of the total number of interval elements between each creative element point and all edge creative element points in the multidimensional creative element network, as well as the degree centrality, betweenness centrality, and proximity centrality of the corresponding creative element point in the multidimensional creative element network, the core position value of each creative element point in the multidimensional creative element network is calculated. Based on the evaluation value of each creative element under the multidimensional evaluation criteria in each initial ad creative and its core position value in the multidimensional creative element network, the effectiveness weight of each creative element in the multidimensional creative element network is calculated. Based on the effectiveness weights of all creative elements in the multidimensional creative element network and the evaluation values ​​of all creative elements under the multidimensional evaluation criteria in each initial ad creative, the evaluation value of each initial ad creative relative to the multidimensional evaluation criteria is calculated.

6. The method for large-scale generation of creative advertising materials according to claim 5, characterized in that, Based on the average degree of the number of margin points between each creative element point and all edge creative element points in the multidimensional creative element network, and the degree centrality, betweenness centrality, and proximity centrality of the corresponding creative element point in the multidimensional creative element network, the core position value of each creative element point in the multidimensional creative element network is calculated, including: When there is a unique creative element point in the multidimensional creative element network that has the largest average degree of the total number of interval element points between all edge creative element points, then the bounding lines of creative element points at all levels of the corresponding creative element point are determined in the multidimensional creative element network, and the core position value of each creative element point in the multidimensional creative element network is determined based on the bounding lines of creative element points at all levels. When there is more than one creative element point in the multidimensional creative element network that has the greatest average degree of the total number of interval element points among all edge creative element points, the bounding line of creative element points at all levels of each creative element point with the greatest average degree is determined. The centrality value of each creative element point in the multidimensional creative element network is calculated based on its degree centrality, betweenness centrality, and proximity centrality. Based on the binding depth between each creative element point with the maximum average degree and all creative element points in the enclosing line of creative element points at all levels, and the product of the centrality values ​​of all creative element points in the enclosing line of creative element points at all levels in the multidimensional creative element network, a comprehensive matrix of the corresponding creative element points is generated. The core position value of each creative element in the multidimensional creative element network is calculated based on the comprehensive matrix of each creative element point with the maximum average degree.

7. The method for large-scale generation of creative advertising materials according to claim 6, characterized in that, The core position value of each creative element in the multidimensional creative element network is calculated based on the comprehensive matrix of each creative element point with the highest average degree, including: Determine the comprehensive similarity between the eigenvector corresponding to the largest eigenvalue of the comprehensive matrix of each creative element point with the maximum average degree and the row vector corresponding to the enclosing line of creative element points at all levels in the comprehensive matrix; Based on the creative element points with the highest average degree, the enclosing line of all levels of creative element points with the highest comprehensive similarity determines the core position value of each creative element point in the multidimensional creative element network.

8. The method for large-scale generation of creative advertising materials according to claim 1, characterized in that, Based on the evaluation values ​​of all initial ad creatives relative to the multidimensional evaluation criteria, the optimal ad creatives are selected, including: Based on the evaluation value of each initial ad creative relative to the multidimensional evaluation criteria, a comprehensive evaluation value for each initial ad creative is determined; The initial ad creative with the highest overall score among all initial ad creatives is considered the optimal ad creative.

9. A system for large-scale generation of creative advertising materials, characterized in that, A method for performing the large-scale generation of creative advertising materials according to any one of claims 1 to 8 includes: The element information combination module is used to obtain a set of creative element expression candidates based on the specific element information of multi-dimensional creative element points, and to combine the creative element expression candidate sets according to rules to obtain multiple initial creative element combinations. The ad creative generation module is used to generate multiple initial ad creatives based on a preset AI model for each initial creative element combination; The element network construction module is used to construct a multi-dimensional creative element network based on the binding depth between multi-dimensional creative element points; The advertising effectiveness evaluation module is used to analyze the evaluation value of each initial advertising creative relative to the multidimensional evaluation criteria based on the multidimensional creative element network. The ad creative selection module is used to select the best ad creatives based on the evaluation values ​​of all initial ad creatives relative to the multidimensional evaluation criteria.

Images