Advertisement creativity and putting strategy intelligent generation system based on multiple platforms

Abstract

The invention discloses a multi-platform-based advertisement creativity and putting strategy intelligent generation system, equipment and medium, multi-source data are integrated through a cross-platform creativity intelligent generation module to generate creativity materials adaptive to different platforms, and a cross-platform collaborative strategy optimization module is used for quantifying a platform collaborative effect and dynamically allocating budget. An advertisement creative life cycle is managed by means of a real-time dynamic optimization module, a strategy is adjusted based on multi-dimensional feedback, and finally, output of each module is fused through a unified decision engine to form a unified decision. The system realizes full-process intelligentization of advertisement originality and putting strategies, quantifies a cross-platform synergistic effect to improve budget distribution efficiency, constructs an advertisement originality life cycle dynamic management mechanism, realizes strategy real-time evolution through multi-dimensional feedback, and visualizes full-link data in combination with a unified decision engine, so that the advertisement putting return on investment is remarkably improved, and the advertisement putting efficiency is improved. The creative manufacturing cost is reduced, the strategy adjustment efficiency is improved, and an efficient and accurate multi-platform putting solution is provided for advertisers.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology and relates to a multi-platform advertising creativity and delivery strategy intelligent generation system, equipment and medium. Background Art

[0002] In the current digital marketing landscape, advertisers face the need to place ads across multiple platforms with diverse characteristics, such as social media, email, and outdoor advertising. However, traditional methods for developing creative ideas and developing delivery strategies have numerous shortcomings. Manually designing ad creatives consumes significant time and labor costs, and the quality and effectiveness of these creatives are difficult to guarantee, making them difficult to adapt quickly to the user characteristics and display rules of different platforms. Furthermore, delivery strategies often rely on experienced marketers, lacking scientific data analysis and intelligent prediction, resulting in poor delivery results and significant resource waste. While some existing ad generation and delivery systems offer some degree of automation, most are limited to a single platform or a limited number of platforms, failing to integrate data and coordinate strategies across multiple platforms. This makes it difficult to meet advertisers' needs for integrated multi-platform marketing. Therefore, there is an urgent need for a system that can intelligently generate ad creatives and develop delivery strategies based on multi-platform data to improve the efficiency and effectiveness of ad delivery. Summary of the Invention

[0003] In order to solve the problems existing in the background technology, the present invention proposes an intelligent generation system for advertising creativity and delivery strategies based on multiple platforms, aiming to solve the problem that existing platforms are unable to achieve data integration and strategy coordination among multiple platforms, and are difficult to meet advertisers' needs for multi-platform integrated marketing.

[0004] The first aspect of the present application provides a multi-platform advertising creative and delivery strategy intelligent generation system, including: a cross-platform creative intelligent generation module, a cross-platform collaborative strategy optimization module, a real-time dynamic optimization module and a unified decision engine module; A cross-platform creative intelligent generation module, which integrates multi-platform user behavior data to construct multi-dimensional feature vectors and generates advertising creative materials adapted to different platforms based on a generative adversarial network combined with the Transformer architecture; A cross-platform synergy strategy optimization module is used to analyze the advertising creative materials to quantify the synergy effects between different advertising delivery platforms and generate a cross-platform budget allocation plan; A real-time dynamic optimization module for managing the advertising creative lifecycle according to the cross-platform budget allocation plan and adjusting the delivery strategy based on multi-dimensional feedback; The unified decision engine module is used to integrate the modules based on the delivery strategy and output a unified decision and visualization interface.

[0005] Optionally, the cross-platform creative intelligent generation module includes a multi-source data fusion feature extraction unit and a generative creative generation unit; the multi-source data fusion feature extraction unit uses an attention mechanism neural network to extract user behavior features of multiple platforms and generate multi-dimensional feature vectors, and the generative creative generation unit is based on a generative adversarial network combined with a Transformer architecture to generate advertising creative materials based on multi-dimensional feature vectors and target platform style features.

[0006] Optionally, the generative creative generation unit includes a content generator, a style transferrer and a diversity enhancer; the content generator inputs target user group characteristics and product attribute characteristics to generate basic creative materials, the style transferr performs style transfer on the basic creative materials according to target platform characteristics, and the diversity enhancer generates creative variants by randomly perturbing the latent vector.

[0007] Optionally, the cross-platform collaboration strategy optimization module includes a platform collaboration effect quantification unit and a dynamic budget allocation unit; the platform collaboration effect quantification unit obtains a cross-platform collaboration factor by calculating the difference ratio between the dual-platform collaboration conversion rate and the single-platform conversion rate, and the dynamic budget allocation unit generates a cross-platform budget allocation plan based on a multi-armed bandit model of reinforcement learning combined with the cross-platform collaboration factor.

[0008] Optionally, the cross-platform collaborative strategy optimization module further includes a personalized strategy generation unit; the personalized strategy generation unit is used to generate a personalized delivery strategy including delivery time, delivery frequency and target population for multiple platforms based on a cross-platform budget allocation plan.

[0009] Optionally, the real-time dynamic optimization module includes an advertising creative lifecycle management unit and multi-dimensional feedback; the advertising creative lifecycle management unit calculates the creative freshness index through a creative decay model; the multi-dimensional feedback adjusts the advertising creative generation model parameters and delivery strategy based on the creative freshness index and the performance data of each platform.

[0010] Optionally, the multi-dimensional feedback dynamically adjusts the cross-platform budget allocation according to the changes in the cross-platform synergy factor, and the budget adjustment amount is calculated based on the adjustment coefficient, the cross-platform synergy factor and the platform average conversion deviation degree; the platform average conversion deviation degree is the difference between the platform conversion rate and the average conversion rate.

[0011] Optionally, the unified decision engine module includes a decision fusion model and a visual decision interface; the decision fusion model forms a unified decision vector by weighted fusion of the outputs of the cross-platform creative intelligent generation module, the cross-platform collaborative strategy optimization module and the real-time dynamic optimization module; the visual decision interface provides a cross-platform creative effect heat map, a strategy synergy effect matrix and a budget dynamic allocation curve.

[0012] Optionally, the intelligent generation system for advertising creativity and delivery strategies based on multiple platforms adopts a distributed architecture, including a data collection layer, a data processing layer, a model layer and an application layer; the data collection layer collects multi-platform user behavior data and advertising effect data, the data processing layer performs data cleaning and feature engineering, the model layer deploys creative generation models, cross-platform collaboration models and dynamic optimization models, and the application layer provides interfaces for connecting to multiple platforms.

[0013] Optionally, the system is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned multi-platform-based advertising creative and delivery strategy intelligent generation system when executing the computer program.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The above-mentioned system for intelligently generating advertising creativity and delivery strategies based on multiple platforms integrates multi-source data through a cross-platform creative intelligent generation module to generate creative materials adapted to different platforms, utilizes a cross-platform collaborative strategy optimization module to quantify platform synergy effects and dynamically allocate budgets, manages the advertising creative life cycle with the help of a real-time dynamic optimization module, and adjusts strategies based on multi-dimensional feedback, and finally integrates the outputs of each module through a unified decision engine to form a unified decision. The system described in the present invention realizes the full-process intelligence of advertising creativity and delivery strategies, quantifies cross-platform synergy effects to improve budget allocation efficiency, constructs a dynamic management mechanism for the advertising creative life cycle, realizes real-time strategy evolution through multi-dimensional feedback, and combines a unified decision engine to visualize full-link data, thereby significantly improving the return on investment of advertising delivery, reducing creative production costs, and improving strategy adjustment efficiency, providing advertisers with efficient and accurate multi-platform delivery solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a schematic diagram of a multi-platform advertising creative and delivery strategy intelligent generation system according to an embodiment of the present invention; DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] In one embodiment, if Figure 1 As shown, a multi-platform advertising creative and delivery strategy intelligent generation system is provided, which is applied in Figure 1Take the example of , for explanation, including the following specific modules: The cross-platform creative intelligent generation module described in this application is used to integrate multi-platform user behavior data to construct a multi-dimensional feature vector, and generate advertising creative materials adapted to different platforms based on a generative adversarial network combined with a Transformer architecture; The cross-platform creative intelligent generation module includes a multi-source data fusion feature extraction unit and a generative creative generation unit. The multi-source data fusion feature extraction unit uses an attention mechanism neural network to extract user behavior features from social platforms, email channels, and outdoor advertising channels and generates a multi-dimensional feature vector. The generative creative generation unit is based on a generative adversarial network combined with a Transformer architecture to generate advertising creative materials based on the multi-dimensional feature vector and the style characteristics of the target platform. The generative creative generation unit includes a content generator, a style transferrer, and a diversity enhancer; the content generator inputs the target user group characteristics and product attribute characteristics to generate basic creative materials, the style transferr performs style transfer on the basic creative materials according to the characteristics of the target platform, and the diversity enhancer generates creative variants by randomly perturbing the latent vector.

[0018] Specifically, in this application, the cross-platform creative intelligent generation module integrates user behavior data from multiple platforms such as social platforms, email channels, outdoor advertising, information services, and short videos by means of multi-dimensional data fusion features. The social platform includes likes and sharing data of the social platform, the email channel includes data such as ad opening rate and dwell time, outdoor advertising includes user behavior data such as the flow of people at the advertisement site and the length of stay of each person, the information service platform includes data such as user needs, and the short video platform includes data such as user preferences. By collecting user behavior data from multiple platforms, a multidimensional feature vector is constructed. Among them, the multidimensional feature vector is constructed by extracting key features using an attention mechanism neural network, and the method of extracting key features is as follows: ; in, Represents the multi-dimensional feature vector integrated by the attention mechanism, which is used to characterize the comprehensive characteristics of multi-platform user behavior. is the social feature vector, is the email feature vector, is the outdoor channel feature vector, each feature vector 、 、 The attention mechanism is input after data preprocessing, and the data preprocessing is to standardize each feature vector data, feature screening and other data preprocessing methods, is the attention weight of the social feature vector, is the attention weight of the email feature vector, is the attention weight of the outdoor channel feature vector, and needs to satisfy + + =1.

[0019] The generative creative generation model is based on the generative adversarial network combined with the Transformer architecture to build a creative generation network: The content generator generates creative materials such as text, images, and videos by inputting target user group characteristics and product attribute characteristics.

[0020] The style migrator performs style migration based on the characteristics of different platforms. The migration formula is as follows: ; in, Indicates the creative material adapted to the target platform generated after style transfer. It is a content feature vector, which includes target user group characteristics (such as age, gender, and interest tags) and product attribute characteristics (such as product functions, selling points, and brand tone). It is input and processed by the content generator in the cross-platform creative intelligent generation module. It is the platform style feature vector, which includes the platform's unique visual style, user interaction habits, such as the like / comment mechanism, content length limit and other dimension feature parameters. The generative adversarial network is a deep learning model consisting of a generator and a discriminator, which is used to fuse content features with platform style features to generate new creative materials.

[0021] The diversity enhancer generates diverse creative variants by randomly perturbing the latent vector to meet the needs of A / B testing.

[0022] In this embodiment, the cross-platform creative intelligent generation module is a core module that integrates multi-platform user behavior data to construct a multi-dimensional feature vector, and uses a generative adversarial network and Transformer architecture to generate advertising creative materials adapted to different platforms. It includes a multi-source data fusion feature extraction unit and a generative creative generation unit. The former uses an attention mechanism neural network to extract user behavior features from social platforms, email channels, outdoor advertising channels, and other channels and generate multi-dimensional feature vectors. The latter, based on a generative adversarial network and Transformer architecture, combines multi-dimensional feature vectors with the style characteristics of the target platform to generate advertising creative materials. It is further subdivided into a content generator, a style transferrer, and a diversity enhancer, responsible for generating basic creative materials, adjusting the style according to the characteristics of the target platform, and generating creative variants through random perturbations. Taking the promotion of a new beauty brand's essence across three typical platforms as an example, this module first collects data from multiple channels, including user discussion keywords about the essence, email open preferences, and outdoor ad viewing time, to generate a multidimensional feature vector containing user profiles, behavioral characteristics, and platform preferences. The content generator then generates basic creative ideas based on this data and product attributes. The style transferer then creates fast-paced short videos with trending background music for short video platforms, sophisticated nine-square grid graphics for social image platforms, and Instagram-style posed notes for lifestyle platforms. The diversity enhancer also generates variations, such as those highlighting sensitive skin and promotional campaigns. The generated creative assets are integrated with the delivery strategy. For example, the sensitive skin version on short video platforms is targeted to users with relevant tags. After delivery, feedback is collected to optimize the feature vector, forming a closed loop.

[0023] This application uses a style migrator to achieve efficient cross-platform adaptation and avoid duplication of work; a multi-dimensional feature vector drives creativity to better suit user preferences and improve advertising effectiveness; a diversity enhancer supports large-scale creative production to meet the needs of multi-scenario delivery; and an intelligent optimization closed loop is constructed in combination with delivery feedback to promote continuous improvement in creative effects, providing strong technical support for the intelligent generation system of multi-platform advertising creativity and delivery strategies.

[0024] A cross-platform synergy strategy optimization module is used to analyze the advertising creative materials to quantify the synergy between different advertising delivery platforms and generate a cross-platform budget allocation plan. The cross-platform synergy strategy optimization module includes a platform synergy effect quantification unit and a dynamic budget allocation unit. The platform synergy effect quantification unit calculates the difference between the dual-platform synergy conversion rate and the single-platform conversion rate to obtain a cross-platform synergy factor. The dynamic budget allocation unit generates a cross-platform budget allocation plan based on a multi-armed bandit model based on reinforcement learning and the cross-platform synergy factor.

[0025] Specifically, the platform synergy effect quantitative model is to define the cross-platform synergy factor , to measure the synergy between platforms i and j: ; in, It represents the cross-platform synergy factor between platform i and platform j, and is used to quantify the effect gain or loss when two platforms are deployed together. The synergy factor between platform i and platform j is independent of the order and only reflects the combined effect of the two. and is the single platform conversion rate, is the advertising conversion rate of platform i, is the advertising conversion rate of platform j, which is the click-through conversion rate and purchase conversion rate, obtained through historical advertising data statistics. is the dual-platform collaborative conversion rate, i and j represent the indexes of the advertising delivery platforms, and can refer to any two different advertising delivery platforms.

[0026] The dynamic budget allocation algorithm is a multi-armed bandit model based on reinforcement learning, which combines synergy factors to optimize the budget: ; in, For all possible strategies Find the maximum value, that is, choose the strategy that can maximize the value of the expression in the square brackets. To express strategy An estimate of the historical average revenue of a platform combination, such as the historical conversion rate or return on investment. , For strategy Number of times it has been executed, is the number of iterations. Increase, The growth is slow, which makes the algorithm gradually converge to the known optimal strategy; if a strategy If is smaller, the exploration factor will increase, encouraging people to try the strategy to obtain more information.

[0027] In this embodiment, the platform synergy effect quantification unit quantifies the synergy effect of different platform combinations by constructing a mathematical model. The core of the model is to calculate the "cross-platform synergy factor". By comparing the conversion rate of collaborative delivery on both platforms , the degree of deviation between the conversion rate of a single platform and the average conversion rate of the platform, which measures the collaborative gain or loss of platforms i and j. For example, when a social platform (platform i) and a short video platform (platform j) collaborate to deliver, if the conversion rates of the single platforms are 5% and 8% respectively, and the collaborative conversion rate is 15%, then the synergy factor =(15%−13%) / 13%≈15.38%, indicating that there is a positive synergy between the two platforms. This unit trains the synergy matrix through historical delivery data to provide a quantitative basis for budget allocation. The process of measuring the collaborative gain or loss between platforms i and j is as follows: First, collect the conversion rate of platform i when it is delivered alone. Conversion rate when platform j is launched alone , and then obtain the comprehensive conversion rate when the two platforms are deployed together By formula Calculating cross-platform synergy factors .like >0, indicating that there is positive synergy between the two platforms, and the synergistic delivery effect is better than the sum of the individual delivery effects; if =0, indicating no synergistic effect; if If <0, there is negative synergy, and the effect of synergistic delivery is inferior to the sum of individual delivery. This process quantifies the synergistic effect of platform combinations, providing data support for cross-platform budget allocation and achieving optimal allocation of advertising resources.

[0028] The dynamic budget allocation unit uses a multi-armed bandit model based on reinforcement learning, taking cross-platform synergy factors as state inputs and searching for the optimal budget allocation strategy through iterative optimization. The multi-armed bandit model simulates the "exploration-exploitation" decision-making process: on the one hand, it explores new platform combinations to discover strategies with high synergy effects (such as trying a combination of social platforms and information service platforms), and on the other hand, it utilizes known efficient strategies (such as continuing to invest in social platforms and short video platforms). , the model is expressed by the formula Dynamically adjust budget allocation, where is the expected return of the strategy, is the number of strategy executions, To increase the number of iterations, ensure that the budget is tilted towards high-synergy, high-return platform combinations.

[0029] For example, take the promotion of an e-commerce brand as an example. It plans to place new product advertisements on social platforms (Platform 1), short video platforms (Platform 2), and information service platforms (Platform 3), with a total budget of 1 million yuan: the synergy effect quantification is to calculate the platform synergy factor through historical data, and obtain =15% (social platform + short video platform, synergistic efficiency), =5% (social platform + information service platform, weak synergy), =8% (short video platform + information service platform, moderate synergy). Dynamic budget allocation initially adopted an even distribution (333,000 yuan per platform). Iterative optimization was performed using a multi-armed bandit model: In the first iteration, the model experimented with increasing the set budget for the social platform + short video platform combination, finding a 12% increase in conversion rate, validating the effectiveness of the synergy factor. In the fifth iteration, based on synergy factors and revenue feedback, the model adjusted the budget to 400,000 yuan for social platforms, 350,000 yuan for short video platforms, and 250,000 yuan for information service platforms, focusing on the high-synergy combination. Compared to the initial even distribution, the optimized cross-platform conversion rate increased by 22%, and the return on investment increased from 1.8% to 2.3%, validating that the budget allocation strategy based on synergy factors can effectively tap into the value of the platform combination and reduce inefficient investment waste.

[0030] A real-time dynamic optimization module is used to manage the advertising creative life cycle according to the cross-platform budget allocation scheme and adjust the delivery strategy based on multi-dimensional feedback. The real-time dynamic optimization module includes an advertising creative life cycle management unit and multi-dimensional feedback; the advertising creative life cycle management unit calculates the creative freshness index through a creative decay model; the multi-dimensional feedback is based on the performance data and creative interaction data of each platform, combined with the creative freshness index to adjust the advertising creative generation model parameters and delivery strategy. The multi-dimensional feedback dynamically adjusts the cross-platform budget allocation according to the changes in the cross-platform synergy factor, and the calculation of the budget adjustment amount is based on the adjustment coefficient, the cross-platform synergy factor and the platform average conversion deviation degree; the platform average conversion deviation degree is the difference between the platform conversion rate and the average conversion rate. The advertising creative generation model dynamically adjusts the advertising creative materials generated by the cross-platform creative intelligent generation module through multi-dimensional feedback, and adjusts the delivery strategy of the advertising creative materials based on the adjustment of the advertising creative materials.

[0031] Specifically, the advertising creative life cycle management is to calculate the creative freshness index in real time by building a creative decay model : ; in, For initial freshness, is the attenuation coefficient, For the delivery time, is the advertising creative life cycle threshold, e x is an exponential function, and e is a constant.

[0032] This multi-dimensional feedback optimizes the advertising delivery system in real time through performance feedback, creative feedback, and cross-platform linkage. Performance feedback involves real-time collection of data such as conversion rates and click-through rates across platforms to update the strategy effectiveness evaluation model. Creative feedback involves analyzing user interaction with creatives (such as comment keywords and dwell time) to optimize the creative generation model. Cross-platform linkage dynamically adjusts cross-platform strategies based on changes in synergy factors, such as: ; in, is the budget adjustment amount of platform i, is the adjustment coefficient, is the platform conversion rate, is the average conversion rate, where .

[0033] In the embodiment of the present application, the real-time dynamic optimization module is the "intelligent adjustment center" of the advertising delivery system, and its core value lies in achieving the adaptive evolution of advertising creativity and delivery strategies through continuous monitoring and feedback mechanisms. This module is based on a cross-platform budget allocation solution and constructs a closed-loop optimization system through two core units: the advertising creative lifecycle management unit quantifies the freshness of creativity, identifies the creative decline inflection point and triggers the update mechanism; the multi-dimensional feedback integrates platform effect data, creative interaction data and cross-platform synergy factors to dynamically adjust budget allocation and creative generation parameters. This dual-engine driven architecture enables the system to maintain the best delivery effect in a rapidly changing market environment, avoiding the lag problem of the traditional "static strategy + periodic adjustment" model.

[0034] Through the advertising creative life cycle and multi-dimensional feedback, the advertising creative delivery strategy is optimized in real time. That is, by quantifying creative decay and multi-dimensional feedback, the system can dynamically adjust the strategy to maximize cross-platform synergy while maintaining the freshness of the creativity, thereby achieving continuous optimization of budget utilization efficiency.

[0035] A unified decision engine module is used to integrate various modules based on the delivery strategy and output a unified decision and visualization interface. The unified decision engine module includes a decision fusion model and a visualization interface. The decision fusion model forms a unified decision vector by weightedly integrating the outputs of the cross-platform creative intelligent generation module, the cross-platform collaborative strategy optimization module, and the real-time dynamic optimization module. The visualization interface provides a cross-platform creative effect heat map, a strategy synergy effect matrix, and a dynamic budget allocation curve.

[0036] Specifically, the decision fusion model integrates the outputs of the cross-platform creative intelligent generation module, the cross-platform collaborative strategy optimization module, and the real-time dynamic optimization module to build a unified decision function. : ; in, is a creative decision vector, such as material type, style preference, is a strategic decision vector, such as platform combination, budget allocation, To optimize decision vectors, such as real-time effect feedback and changes in synergy factors, is the weight coefficient of the creative decision vector, is the weight coefficient of the strategy decision vector, To optimize the weight coefficient of the decision vector, the weight coefficient 、 、 , dynamically adjusted based on historical performance data and business goals, for example: when launching a new product Increase, mature product maintenance Increase.

[0037] The visual decision interface provides visualization components such as a cross-platform creative effect heat map, a strategy synergy matrix, and a budget dynamic allocation curve, and supports manual fine-tuning of decision parameters.

[0038] It is important to note that the visual decision-making interface supports decision-making through three core components: Cross-platform creative effect heat map: Displays indicators such as the click-through rate and conversion rate of creatives on each platform in shades of color, allowing intuitive identification of high- and low-efficiency creatives; Strategic synergy matrix: Presents synergy factors between platforms in matrix form , assisting in selecting the optimal platform combination; dynamic budget allocation curve: displays the budget allocation ratio and historical adjustment trajectory of each platform in real time, and supports manual fine-tuning and verification.

[0039] In this application, the unified decision engine is the "brain" of the advertising intelligent delivery system. Its core function is to deeply integrate the outputs of the three modules of creative generation, strategy optimization and dynamic feedback to form an executable unified decision, and present the decision basis and results through a visual interface. This architecture solves the problem of "multi-module separation and decision fragmentation" in traditional advertising systems, and realizes full-link data-driven decision-making through a weighted fusion mechanism. For example, when the cross-platform creative intelligent generation module recommends type A video material, the cross-platform collaborative strategy optimization module recommends focusing on short video platforms, and the real-time dynamic optimization module prompts that the budget allocation needs to be adjusted, the decision engine will comprehensively calculate the output weights of each module and generate the optimal combination strategy to ensure the consistency and efficiency of advertising delivery.

[0040] In this embodiment, through data-driven fusion decision-making and intuitive visualization, the system can not only generate efficient strategies, but also help decision makers understand "why such decisions are made", realize the organic combination of intelligent decision-making and human experience, and ultimately improve the accuracy and controllability of advertising delivery. Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment systems can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned systems. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM). Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0041] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A multi-platform advertising creative and delivery strategy intelligent generation system, characterized by: include: Cross-platform creative intelligent generation module, cross-platform collaborative strategy optimization module, real-time dynamic optimization module and unified decision engine module; A cross-platform creative intelligent generation module, which integrates multi-platform user behavior data to construct multi-dimensional feature vectors and generates advertising creative materials adapted to different platforms based on a generative adversarial network combined with the Transformer architecture; A cross-platform synergy strategy optimization module is used to analyze the advertising creative materials to quantify the synergy effects between different advertising delivery platforms and generate a cross-platform budget allocation plan; A real-time dynamic optimization module for managing the advertising creative lifecycle according to the cross-platform budget allocation scheme and adjusting the delivery strategy based on multi-dimensional feedback; The unified decision engine module is used to integrate the modules based on the delivery strategy and output a unified decision and visualization interface.

2. The multi-platform advertising creative and delivery strategy intelligent generation system according to claim 1, characterized in that: The cross-platform creative intelligent generation module includes a multi-source data fusion feature extraction unit and a generative creative generation unit; the multi-source data fusion feature extraction unit uses an attention mechanism neural network to extract user behavior features of multiple platforms and generate multi-dimensional feature vectors. The generative creative generation unit is based on a generative adversarial network combined with a Transformer architecture to generate advertising creative materials based on multi-dimensional feature vectors and the style characteristics of the target platform.

3. The multi-platform advertising creative and delivery strategy intelligent generation system according to claim 2, characterized in that: The generative creative generation unit includes a content generator, a style transferr and a diversity enhancer; the content generator inputs target user group characteristics and product attribute characteristics to generate basic creative materials, the style transferr performs style transfer on the basic creative materials according to the characteristics of the target platform, and the diversity enhancer generates creative variants by randomly perturbing the latent vector.

4. The multi-platform advertising creative and delivery strategy intelligent generation system according to claim 1, characterized in that: The cross-platform collaboration strategy optimization module includes a platform collaboration effect quantification unit and a dynamic budget allocation unit; the platform collaboration effect quantification unit obtains a cross-platform collaboration factor by calculating the difference ratio between the dual-platform collaboration conversion rate and the single-platform conversion rate, and the dynamic budget allocation unit generates a cross-platform budget allocation plan based on a multi-armed bandit model of reinforcement learning combined with the cross-platform collaboration factor.

5. The multi-platform advertising creative and delivery strategy intelligent generation system according to claim 4, characterized in that: The cross-platform collaborative strategy optimization module also includes a personalized strategy generation unit; the personalized strategy generation unit is used to generate a personalized delivery strategy including delivery time, delivery frequency and target population for multiple platforms based on the cross-platform budget allocation plan.

6. The multi-platform advertising creative and delivery strategy intelligent generation system according to claim 1, characterized in that: The real-time dynamic optimization module includes an advertising creative lifecycle management unit and multi-dimensional feedback; the advertising creative lifecycle management unit calculates the creative freshness index through a creative decay model; the multi-dimensional feedback adjusts the advertising creative generation model parameters and delivery strategy based on the creative freshness index and the performance data of each platform.

7. The multi-platform advertising creative and delivery strategy intelligent generation system according to claim 6, characterized in that: The multi-dimensional feedback dynamically adjusts the cross-platform budget allocation according to the changes in the cross-platform synergy factor. The budget adjustment amount is calculated based on the adjustment coefficient, the cross-platform synergy factor and the platform average conversion deviation degree; the platform average conversion deviation degree is the difference between the platform conversion rate and the average conversion rate.

8. The multi-platform advertising creative and delivery strategy intelligent generation system according to claim 1, characterized in that: The unified decision engine module includes a decision fusion model and a visual decision interface; the decision fusion model forms a unified decision vector by weighted fusion of the outputs of the cross-platform creative intelligent generation module, the cross-platform collaborative strategy optimization module and the real-time dynamic optimization module; the visual decision interface provides a cross-platform creative effect heat map, a strategy synergy effect matrix and a budget dynamic allocation curve.

9. The multi-platform advertising creative and delivery strategy intelligent generation system according to any one of claims 1 to 8, characterized in that: The system adopts a distributed architecture, including a data collection layer, a data processing layer, a model layer and an application layer; the data collection layer collects multi-platform user behavior data and advertising effect data, the data processing layer performs data cleaning and feature engineering, the model layer deploys creative generation models, cross-platform collaboration models and dynamic optimization models, and the application layer provides interfaces for connecting to multiple platforms.

Images