Ad creative dynamic evaluation and intelligent decision system based on multi-source data fusion

Abstract

The application discloses an advertisement creative dynamic evaluation and intelligent decision system based on multi-source data fusion, and relates to the technical field of advertisement design. The system comprises a state perception and graphing module, an online decision module, an execution feedback module and a collaborative evolution updating module. The state perception and graphing module is used for acquiring real-time interactive behavior and external environment data, mapping the data to a dynamic creative feature graph and outputting a graph structured state vector. The online decision module takes the state vector as input, calls an incremental learning decision model to calculate expected performance values and decision uncertainty values of each candidate creative combination, and generates a creative selection instruction according to the decision uncertainty values through a strategy function. The execution feedback module outputs the instruction and receives corresponding actual performance data. The collaborative evolution updating module synchronously updates the weights of related nodes and edges in the dynamic creative feature graph and the internal parameters of the decision model according to the performance data, the selection instruction and the state vector. The application realizes real-time evaluation, intelligent decision and collaborative self-evolution of advertisement creatives in a dynamic delivery environment.

Description

Technical Field

[0001] This invention relates to the field of advertising design technology, specifically to a dynamic evaluation and intelligent decision-making system for advertising creativity based on multi-source data fusion. Background Technology

[0002] Intelligent evaluation and decision-making of advertising creatives is a core development direction in the field of digital marketing. It aims to improve the return on marketing investment by accurately predicting and optimizing advertising effectiveness through the analysis and modeling of massive amounts of data. In recent years, with the advancement of big data and machine learning technologies, advertising strategies have gradually evolved towards data-driven, dynamically adjusted, and intelligent approaches.

[0003] Existing technologies typically construct static creative feature knowledge bases or graphs and assign weights or scores to creative elements based on historical performance data (such as click-through rates and conversion rates). When faced with a new target audience, the system calculates the semantic similarity between the target features and creative elements, or directly selects the creative combination with the highest historical score for delivery. This type of method can leverage historical experience to a certain extent to achieve an initial match between advertising creatives and target audiences, possessing static recommendation capabilities based on historical data.

[0004] However, advertising is essentially a sequential decision-making process closely interacting with the dynamic market environment and real-time user feedback. Existing methods based on static knowledge bases and post-event historical scoring have the following interconnected fundamental limitations: First, their decision-making relies on lagging historical performance data accumulated over a period of time, failing to perceive and respond in real-time to rapid changes in current user preferences, competitive environment, or external hot events. Second, their decision-making mechanism is essentially a "one-off" offline matching or retrieval, lacking a closed loop that enables continuous learning, proactive exploration, and strategy iteration during the advertising process. This leads to the system becoming rigid when facing new user groups, new market scenarios, or new creative elements, unable to validate hypotheses, discover potential high-quality creatives, or promptly avoid ineffective strategies through real-time interaction.

[0005] Existing methods struggle to achieve adaptive optimization and maximize long-term returns for advertising creatives during long-term, dynamic campaigns, especially in rapidly changing market environments where the stability and growth of campaign effectiveness face bottlenecks. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a dynamic evaluation and intelligent decision-making system for advertising creatives based on multi-source data fusion.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows:

[0008] This invention discloses a dynamic evaluation and intelligent decision-making system for advertising creatives based on multi-source data fusion, comprising:

[0009] The state awareness and graphing module is used to acquire real-time user interaction behavior data and external environmental event data, and map them to a pre-stored dynamic creative feature graph for matching and mapping processing, outputting a graph-structured state vector.

[0010] The online decision-making module is used to take the graph structured state vector as input, call a pre-trained incremental learning decision model, calculate the expected performance value and decision uncertainty value of multiple candidate creative combinations sampled from the dynamic creative feature graph; and generate creative selection instructions based on the decision uncertainty value through a preset strategy function; wherein, when the decision uncertainty value is higher than a first preset threshold, the strategy function prioritizes creative combinations with low node weights but high correlation edge strength;

[0011] The execution feedback module is used to output the creative selection instruction to the ad delivery execution terminal and receive the actual performance data corresponding to the creative selection instruction from the ad delivery execution terminal.

[0012] The collaborative evolution update module is used to synchronously update the weights and edge strengths of relevant nodes in the dynamic creative feature graph based on the actual performance data, the corresponding creative selection instructions, and the graph structured state vector, and to update the internal parameters of the incremental learning decision model.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0014] 1. This invention maps real-time user interaction behavior and external environmental events to a dynamic creative feature map, generating a structured state vector that reflects the current market environment and user attention focus. This process enables the system to transcend traditional methods that rely on single or lagging historical data, incorporating more accurate and comprehensive contextual information during decision-making, thereby providing a highly relevant environmental perception foundation for subsequent intelligent decision-making.

[0015] 2. This invention achieves an intelligent trade-off between exploration and utilization by invoking a pre-trained incremental learning decision model and dynamically switching strategy functions based on decision uncertainty values. Especially when uncertainty is high, the strategy of prioritizing creative combinations with low node weights but high edge correlation allows the system to proactively discover and validate novel creative combinations with insufficient historical performance data but tightly interconnected internal elements and potential synergistic effects. This enables the system to effectively avoid getting trapped in local optima during the initial launch phase or when the market environment changes drastically, systematically broadening the path to discovering high-quality creative ideas.

[0016] 3. This invention utilizes performance data generated from actual campaign deployments to synchronously drive the weight updates of nodes and edges in the dynamic creative feature graph and adjust the internal parameters of the decision-making model, constructing a closed loop of knowledge and strategy co-evolution. This allows the system's knowledge base (graph) and decision logic (model) to continuously self-optimize with each campaign feedback, mutually reinforcing each other. This improves the ability to maintain strategy adaptability and performance stability in long-term, dynamic advertising campaigns, overcoming the performance degradation problem of static models caused by environmental changes. Attached Figure Description

[0017] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein:

[0018] Figure 1 This is a system module connection diagram of the present invention;

[0019] Figure 2 This is a flowchart of the workflow steps of the present invention;

[0020] Figure 3 This is a flowchart of the matching mapping process of the present invention. Detailed Implementation

[0021] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0022] In existing technologies, ad creative optimization largely relies on historical data matching or offline content generation, making it difficult to cope with real-time changes in the market environment and user feedback. Existing systems lack a closed-loop online decision-making mechanism, failing to dynamically evaluate creative effectiveness and adjust delivery strategies based on real-time interactive data. Especially in the initial stages of a campaign or during sudden market changes, existing static recommendation models cannot effectively balance the conflicting decisions of "exploring new creatives" and "utilizing known creatives," resulting in difficulties in continuously optimizing long-term campaign performance.

[0023] To address the aforementioned issues, this application deeply integrates a creative element knowledge graph with a reinforcement learning decision model. It constructs a state vector of the advertising environment through dynamic perception of real-time data and adaptively selects creative delivery strategies based on decision uncertainty. Furthermore, it synchronously updates the node weights and association strengths of the knowledge graph using actual performance data after delivery, while simultaneously optimizing the internal parameters of the decision model, forming a continuously evolving intelligent closed loop.

[0024] After introducing the basic concept of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Example:

[0026] like Figure 1 As shown, the advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion includes:

[0027] The state awareness and graphing module acquires real-time user interaction data and external environmental event data, maps them to a pre-stored dynamic creative feature graph for matching and mapping processing, and outputs a graph-structured state vector. The dynamic creative feature graph is essentially a graph structure stored in an in-memory database (such as Neo4j), where nodes are vectorized creative genes (e.g., visual style vectors and copywriting theme vectors extracted through deep neural networks), and edges represent the co-occurrence or semantic association strength between creative genes. Matching and mapping processing involves using natural language processing techniques to parse keywords from event data and calculate similarity with the text labels of graph nodes; simultaneously, through a rule engine or simple model, user behavior (e.g., clicking on a certain area in an advertisement) is mapped to specific creative gene nodes representing the visual elements of that area.

[0028] The initial construction of the dynamic creative feature map includes: extracting visual and textual creative elements from a historical ad creative library; extracting visual feature vectors using a pre-trained convolutional neural network (such as ResNet-50); and extracting textual semantic vectors using a Sentence-BERT model, which serve as nodes in the map. Edges between nodes are constructed based on the co-occurrence frequency of creative elements in historical ads, with the initial association strength set to a normalized value of the co-occurrence frequency. Initial node weights can be set to a normalized value of historical click-through rates, or uniformly initialized to a median value (such as 0.5). The map is stored in a graph database (such as Neo4j) and supports real-time querying and updating.

[0029] The online decision-making module takes a graph-structured state vector as input, calls a pre-trained incremental learning decision model, and calculates the expected performance value and decision uncertainty value of multiple candidate creative combinations sampled from the dynamic creative feature graph. Based on the decision uncertainty value, it generates creative selection instructions through a preset policy function. The preset policy function dynamically switches between exploration and utilization strategies based on the comparison between the decision uncertainty value and a preset threshold. When the decision uncertainty value is higher than a first preset threshold, the policy function prioritizes creative combinations with low node weights but high edge strength. The incremental learning decision model includes, but is not limited to, a context-based multi-armed gambling machine model.

[0030] The pre-trained model for the incremental learning decision model is obtained through offline training using publicly available advertising datasets (such as Criteo and Avazu). Model weights are initialized using the Xavier normal distribution initialization method, and biases are initialized to zero. During pre-training, the input features are historical advertising context vectors, and the labels are click / conversion signals. The mean squared error loss function is used, and the model is trained to convergence using the Adam optimizer. The trained model parameters are loaded as the initial decision model for the system, supporting subsequent online incremental learning.

[0031] The execution feedback module is used to output creative selection instructions to the ad delivery execution terminal and receive the actual performance data corresponding to the creative selection instructions from the ad delivery execution terminal.

[0032] The co-evolution update module is used to synchronously update the weights and edge strengths of relevant nodes in the dynamic creative feature map based on actual performance data, corresponding creative selection instructions, and graph structured state vectors, and to update the internal parameters of the incremental learning decision model.

[0033] like Figure 2 The diagram shows the workflow steps of this invention. The specific workflow is as follows: During system initialization, a random strategy or an initialization method based on historical data is used. The state awareness and graphing module first continuously acquires real-time data streams from the data source using stream processing technology, matching and mapping each piece of raw data with a dynamic creative feature graph pre-stored in the distributed storage module. After matching, based on the matching strength, relevant nodes and their associated edges are temporarily weighted and marked in the memory copy of the dynamic creative feature graph, thereby simulating the "instantaneous impact" of the current environment on the local structure of the graph. Subsequently, the structural information of the weighted local subgraph (such as key node IDs, temporary weights, and edge strengths) is encoded into a fixed-dimensional graph structured state vector, serving as a digital snapshot of the current advertising environment.

[0034] The online decision-making module receives the structured state vector of the creative feature map as model input. Based on a preset sampling strategy (such as probability sampling based on node weights), it generates multiple candidate creative combinations in real time from the dynamic creative feature map, and calculates two core outputs in parallel: the expected performance value (a scalar predicting the expected click-through rate or conversion rate of the creative combination in the current state) and the decision uncertainty value (such as the prediction variance obtained through a Bayesian neural network, or the divergence degree obtained through ensemble learning). The preset strategy function makes a decision based on the uncertainty value: if the uncertainty value is higher than a first preset threshold, it indicates that the model's understanding of the current environment is insufficient, and "exploration" is required. In this case, the strategy function will select creative combinations with lower average node weights (representing mediocre or novel historical performance) but higher average strength of the edges connecting internal nodes (representing good synergy and structural stability within the combination), and prioritize one of them. Finally, a structured creative selection instruction is output, including the identifier of the target combination and the delivery parameters.

[0035] The execution feedback module sends creative selection instructions to the ad execution end (such as an ad exchange platform) in near real-time. After a campaign period (such as several hours) ends, the execution feedback module receives aggregated actual performance data (such as impressions, clicks, conversion costs, etc.) corresponding to this instruction from the ad execution end.

[0036] The co-evolutionary update module receives actual performance data, corresponding historical selection instructions, and the graph-structured state vector used in decision-making from the feedback module. Its update process is synchronous and cooperative, including:

[0037] 1. Update Dynamic Creative Feature Map: The system identifies the specific creative gene nodes involved in the selection command. Based on the comparison between actual performance data and preset benchmarks, a reward signal is generated. For the dynamic creative feature map, this reward signal is used to adjust the long-term weights of relevant nodes (e.g., updating weight values ​​using a time-decay weighted average algorithm) and the strength of the edges between nodes (e.g., if the current campaign performs well, the strength of the edges between all nodes in this combination is increased).

[0038] 2. Update the internal parameters of the incremental learning decision model: At the same time, the same reward signal mentioned above, together with the historical state vector, constitutes a training sample, which is used to fine-tune the incremental learning model in the online decision module (e.g., update the neural network weights through stochastic gradient descent) to make its future predictions more accurate.

[0039] This application enables online, autonomous, and continuous optimization of advertising creative delivery strategies. It can dynamically adapt to changes in the market environment and user preferences, intelligently balancing creative exploration and utilization, thereby significantly improving long-term advertising performance (such as overall return on investment (ROI)). This application not only outputs decision results, but its internal knowledge representation and decision logic also continuously evolve with data accumulation, effectively solving the cold start problem and model drift problem, and enhancing the system's robustness and adaptability in complex and ever-changing delivery environments.

[0040] like Figure 3 The diagram shown is a flowchart of the matching mapping process of the present invention. This application further proposes that the specific steps for obtaining real-time user interaction behavior data and external environmental event data, and mapping them to a pre-stored dynamic creative feature map for matching mapping processing, include:

[0041] Obtain the pre-stored dynamic creative feature map, which is constructed with creative gene vectors as nodes. The visual element gene vectors (such as the fc7 layer features extracted by the VGG16 network) are associated with the standardized spatial coordinate region of the advertising canvas.

[0042] When inputted from real-time interactive behavior data (e.g., generated by a front-end eye-tracking SDK or mouse trajectory heatmap), the system first parses and generates a heatmap of the user's gaze dwell time in the advertising canvas area. This heatmap is a two-dimensional matrix matched to the canvas resolution, where the value of each element represents the cumulative gaze duration or interaction event density of the corresponding pixel area within a preset time window (e.g., the most recent 10 seconds).

[0043] The threshold segmentation algorithm in image processing is used to extract the coordinate sequence of regions whose intensity exceeds a preset intensity threshold from the heatmap data of the gaze-attention heatmap. This preset intensity threshold can be set through historical data analysis, for example, by taking the 75th percentile of the intensity values ​​in all users' historical heatmap data, with an empirical range of (0.3, 0.7) (after normalization).

[0044] The system matches and associates the region coordinate sequences with the creative gene vectors representing visual elements stored in the dynamic creative feature map. The matching process is achieved by calculating the intersection-over-union (IoU) ratio between the coordinate sequences and the preset coordinate regions of the creative gene vectors. A successful association is determined when the IoU exceeds a matching threshold (e.g., 0.5). A coordinate mapping table is then used to map the heatmap regions to preset visual gene nodes. This mapping table is constructed based on the relationship between the advertising canvas resolution and the gene vector coordinates.

[0045] Based on the matching and association results, temporary enhancement weight coefficients are generated for the relevant nodes and associated edges, and these enhancement weight coefficients are encoded into the graph structured state vector. The calculation formula is:

[0046]

[0047] in, To match the total area of ​​the high-attention region, This represents the total area of ​​the advertising canvas. Characterizes attention concentration; This represents the average length of stay in the area. The baseline dwell time (e.g., 1 second); This is a modulating factor used to control the overall scale of the enhancement magnitude. ∈[0.1,0.5];log( The function is used to smooth out the effect of duration.

[0048] The system will calculate the temporary enhancement weight coefficients. Together with the corresponding node identifiers and associated edge identifiers, they are encoded into a specific dimension range of the graph structured state vector.

[0049] The construction and incremental updating of the dynamic creative feature map is a continuous learning process. The initial creative feature map can be trained from a historical ad creative library. Textual features are extracted using models such as Word2Vec and Sentence-BERT, and visual features are extracted using pre-trained convolutional neural networks (such as ResNet-50). The feature vectors are used as nodes. The edges between nodes and the initial association strength can be calculated by the co-occurrence frequency of creative elements in historical ads, or obtained by training a graph neural network on ad performance data. The online incremental update of the map is driven by a co-evolutionary update module. It uses actual campaign performance data (such as click-through rate, CTR) as a reward signal, updates node weights through a time-decay-based weighted average algorithm, and adjusts the strength of association edges through an optimization method similar to stochastic gradient descent, ensuring that the map knowledge continues to evolve with the campaign process.

[0050] This application enables online decision-making models to explicitly consider the spatial distribution of real-time user attention when evaluating creative combinations, thus favoring creative combinations that include currently highly attention-grabbing visual elements and improving the accuracy of matching advertising creatives with users' immediate interests. It also ensures a balance between capturing fleeting trends and maintaining long-term stable knowledge.

[0051] like Figure 3 The diagram shown is a flowchart of the matching mapping process of the present invention. This application further proposes that the specific steps for obtaining real-time user interaction behavior data and external environmental event data, and mapping them to a pre-stored dynamic creative feature map for matching mapping processing, include:

[0052] The system extracts keywords or sentiment tags from external environmental event data. Specifically, it continuously acquires external environmental event data streams from pre-defined news aggregation APIs, social media feeds, or search engine trends. These data streams typically include text titles, summaries, and publication timestamps. The event text is then processed using a pre-trained natural language processing model. For example, a BERT-based sequence labeling model can be used to extract key entities and noun phrases as keywords, while a RoBERTa-based sentiment analysis model outputs the overall sentiment tag of the text (e.g., positive, negative, neutral, and their confidence level).

[0053] The system performs semantic similarity matching between keywords or sentiment tags and predefined creative gene vectors representing text themes or sentiments in a dynamic creative feature graph. Creative gene vectors are typically representative vectors obtained by encoding and clustering massive amounts of advertising copy and brand slogans using sentence encoding models such as Sentence-BERT. Each vector is associated with one or more semantic tags. For keywords, the system encodes them as word vectors (e.g., using Word2Vec or directly using Sentence-BERT to encode short sentences), and then calculates their cosine similarity with the creative gene vectors of each text. For sentiment tags, the system performs consistency comparison with the sentiment attributes associated with the graph nodes.

[0054] The system sets a preset matching threshold (e.g., cosine similarity greater than 0.7, or sentiment consistency of positive / negative with a confidence level greater than 0.8) to determine the significance of the association. The range of the preset matching threshold can be set based on the similarity distribution of effective associations in historical data, with empirical values ​​typically between 0.65 and 0.75.

[0055] Nodes in the dynamic creative feature map whose matching degree exceeds a preset matching threshold are marked as nodes affected by the current external event, and their influence factors are set accordingly. Encode the structured state vector into the graph.

[0056] To quantify the degree of impact, the impact factor... The calculation can comprehensively consider the matching score and the event's popularity (such as the frequency of mentions on social media). In one embodiment, the calculation formula is as follows:

[0057]

[0058] in, For the normalized semantic matching score (e.g., cosine similarity); This refers to the number of times the event has been mentioned or its popularity index within the most recent time window; This is the baseline heat value, used for standardization; These are weighting coefficients. ∈[0,1], for example, take 0.6, to balance the influence of semantic relevance and event popularity.

[0059] Ultimately, the system assigns the identifiers of these nodes and their corresponding influence factors. The value is encoded into another specific dimension interval of the graph structured state vector.

[0060] This application dynamically links the semantic information of external environmental events to a creative feature map and quantifies it into influencing factors encoded into a state vector. This enables the online decision-making module to perceive and respond to changes in the macro-social environment. The decision-making model can therefore tend to select creative combinations that are semantically relevant to current hot topics or align with mainstream sentiment, thereby improving the environmental adaptability and timeliness of advertising content. This mechanism allows advertising strategies to leverage social hot topics, avoiding the placement of inappropriate advertising content in negative public opinion environments, thus enhancing brand safety and communication effectiveness.

[0061] This application further proposes that the incremental learning decision model invoked is preferably a contextual multi-armed slot machine model. Its technical principle lies in formalizing the dynamic decision-making problem of advertising creatives into a sequential decision-making task: at each decision time slot (e.g., every half hour or for each batch of new user traffic), the system needs to select one creative (arm) from a set of candidate creatives for delivery, and evaluate the quality of the selection based on user feedback (rewards) obtained immediately or later, with the goal of maximizing long-term cumulative returns (such as total clicks or conversion value). The contextual multi-armed slot machine model specifically includes:

[0062] The "arms" in the contextual multi-armed gambling machine model are not fixed, but dynamically generated in each decision slot. Each arm of the contextual multi-armed gambling machine model strictly corresponds to a candidate creative combination sampled in real time from the dynamic creative feature map. The candidate creative combination is generated from the dynamic creative feature map based on the node weight (representing historical performance), so that high-weight nodes have a higher probability of being selected. At the same time, it is generated by sampling the strength of the associated edges, and tends to combine nodes with high associated edge strength in the graph to take advantage of the synergistic effect between creative elements.

[0063] When constructing the combination, for each decision slot, the context input of the contextual multi-armed gambler model is a graph-structured state vector. This state vector integrates information such as real-time user attention and the influence of external events, providing the model with a feature representation of the current decision-making environment.

[0064] When initializing the model or encountering a completely new creative combination (arm), it is necessary to set the initial expected return. The prior distribution parameters of the returns for each arm in the contextual multi-armed gambling machine model are initialized based on the current weights and edge strengths of the creative gene vector nodes corresponding to the current arm in the dynamic creative feature graph. Specifically, for any new arm (corresponding to a set of creative gene nodes), the model calculates the initial expected return value of the arm based on the current weights of these nodes in the graph (representing the historical average performance of the element) and the average strength of the edge strengths between nodes within the combination (representing the internal synergy of the combination). For example, the mean of the initial return prior distribution is... It can be set in the following ways:

[0065]

[0066] in, It is the average weight of the nodes contained in the arm (normalized to the [0,1] interval). It is the average strength of all related edges between these nodes in the graph (normalized to the [0,1] interval). It is a weighting factor; for example, setting it to 0.7 indicates that the individual historical performance of the node is more trusted.

[0067] Variance of the return distribution It can be set as a... The inversely proportional quantity reflects the uncertainty brought about by the sufficiency of historical data.

[0068] Training a context-based multi-armed gambling machine model is a continuous incremental learning process. Training conditions include: a learning rate typically set to a small value (e.g., 0.01 to 0.1) to accommodate the non-stationarity of the data; a neural network architecture with 3 fully connected layers, 1000 training epochs, an Adam optimizer, pre-training on publicly available advertising datasets (e.g., Avazu, Criteo), 5-fold cross-validation, and a balance between exploration and exploitation using strategies such as Thompson Sampling or UCB (upper confidence bound). Whenever a creative combination is selected and delivered, the collected actual performance data (e.g., click-through rate) is converted into a scalar reward, which, along with the then-input context vector, constitutes a training sample. The model then updates its internal parameters based on this sample (e.g., updating the weight vector in a linear context-based gambling machine, or updating the posterior distribution in Bayesian linear regression), thereby achieving rapid absorption of new knowledge and continuous improvement in future predictions.

[0069] By employing a contextual multi-armed gambling machine model and innovatively utilizing dynamic graph knowledge for arm initialization and sampling, this decision-making mechanism effectively addresses the highly dynamic and uncertain nature of the advertising environment. It not only leverages rich contextual information to make conditional decisions, but its incremental learning characteristics also ensure that the strategy adapts rapidly to real-time feedback. This allows the system to establish a quantifiable, probabilistic model-based balance between exploring novel ideas and utilizing mature ones, thereby achieving more stable and superior cumulative advertising results in long-term operation.

[0070] This application further proposes that, based on the decision uncertainty value, the specific steps for generating creative selection instructions through a preset strategy function include setting two key thresholds:

[0071] First preset threshold With the second preset threshold And satisfy These two thresholds can be set based on the distribution of the model's prediction confidence on a historical validation set. For example, the 80th percentile of the sorted decision uncertainty values ​​can be used as the first preset threshold. The 20th percentile is taken as the second preset threshold. The value range is (0.3, 0.7).

[0072] When the decision uncertainty value exceeds a first preset threshold, the strategy function adopts an exploratory strategy, which includes: calculating the average weight of the creative gene vector nodes contained in each candidate creative combination. The weight of a node reflects its long-term performance in the dynamic graph; a high weight indicates consistently excellent historical performance. The goal of the exploration is to discover potential high-quality ideas; therefore, the system sets a preset weight threshold. (For example, take the median or mean of all node weights). Prioritize nodes with an average weight lower than a preset weight threshold (i.e.,...). Secondly, the system sets a preset correlation strength threshold. (For example, take the 75th percentile of all edge strength values), and further calculate the average strength of the associated edges between nodes within the candidate combinations that meet the weighting conditions. Higher than the preset association strength threshold (i.e. (Candidate creative combinations)

[0073] When the decision uncertainty value is lower than the second preset threshold This indicates that the model is highly confident in its judgment of the current environment. At this point, the combination of candidate ideas with the highest expected performance value is selected to obtain the most predictable immediate return.

[0074] This application, through the design of a strategy function, enables the system's decision-making behavior to adaptively switch based on the model's own cognitive state (uncertainty). This improves the efficiency and systematic nature of discovering the value of novel creative ideas. It allows the system to continuously maintain an adaptive balance between exploration and utilization in a changing environment, effectively avoiding short-term revenue losses due to over-exploration or creative aging and strategy rigidity due to over-utilization. This provides a methodological guarantee for the long-term, stable improvement of advertising effectiveness.

[0075] This application further proposes that the specific steps for synchronously updating the weights of relevant nodes and the strength of associated edges in the dynamic creative feature graph include:

[0076] Upon receiving actual performance data (e.g., click-through rate, CTR), the co-evolutionary update module first compares this actual performance data with a preset benchmark performance value. Based on the comparison result, it generates weight adjustment coefficients. The preset benchmark performance value can be the historical average performance of the ad placement or similar products, an industry benchmark, or a dynamic moving average. Weighting adjustment coefficient. It can be defined as the ratio of actual performance to baseline performance minus 1. Alternatively, a piecewise function mapping can be used to provide a positive adjustment for excellent performance (e.g., exceeding the baseline by 20%) and a negative adjustment for poor performance.

[0077] The system then identifies the creative gene vector nodes involved in the creative selection instruction, that is, the graph nodes corresponding to all creative elements contained in the delivered advertisement.

[0078] The system employs a time-decay-based weighted update algorithm, combined with a weight adjustment coefficient, to calculate the updated node weights for the current weight values ​​of the creative gene vector nodes. This updated weights are then written back to the dynamic creative feature map. The system uses a time-decay-based weighted update algorithm to calculate the updated node weights. The formula is as follows:

[0079]

[0080] in, This is the node weight before the update; These are the weight adjustment coefficients generated this time; It is a time decay function, whose value increases with the time since the last weight update of the node. up to the current time The interval increases and decreases. For example, ,in This is the decay rate parameter. For example, setting it to 0.1 means that the effect is halved approximately every 10 time units. It is the learning rate, which controls the size of the step size in a single update. It is set through grid search or experience, and is usually set to a small positive number (such as between 0.05 and 0.2). Is with The adaptive factor of linkage, among which, .

[0081] Using this formula, a long-unupdated ( The weight of nodes with smaller values ​​is affected by the new evidence. The impact is relatively small, mainly maintaining historical values; while a value that has been frequently updated recently ( Nodes with weights close to 1 have weights that more sensitively reflect recent performance trends. Finally, the updated node weights are calculated. It is written back to the dynamic creative feature map, replacing the original weight values.

[0082] Employing a time-decay-based weighted update algorithm, the weight evolution of nodes in the dynamic creative feature map possesses both memory and adaptability. This allows node weights to smoothly track their long-term performance trends, exhibiting robustness to short-term noisy data while gradually incorporating new and valid evidence, adapting to changes in the value of creative elements over time and in the environment. It avoids the update lag or excessive fluctuations that may arise from simple averaging or fixed-window statistics in traditional methods, thus improving overall stability in non-stationary environments.

[0083] This application further proposes that the specific steps for synchronously updating the weights of relevant nodes and the strength of associated edges in the dynamic creative feature graph also include:

[0084] When evaluating the effectiveness of a campaign, the collaborative evolution update module first performs a binary judgment based on actual performance data to determine whether the campaign was effective in the current decision-making time slot. This judgment is based on comparing key performance indicators (such as click-through rate and conversion rate) with a dynamic effectiveness threshold. The effectiveness threshold can be a moving average of recent historical performance data or a fixed benchmark set based on business objectives (e.g., click-through rate > 0.5%). If the actual performance exceeds this effectiveness threshold, the campaign is considered effective.

[0085] If deemed valid, the system performs edge enhancement. This enhances the strength of all associated edges between the creative gene vector nodes involved in the creative selection instruction. The enhancement magnitude is then determined. This can be positively correlated with the relative excellence of this campaign. Where E represents the performance index ( These are actual performance metrics (such as click-through rate and conversion rate). (This is a validity threshold used to determine whether the campaign is effective.) This is the enhancement coefficient, usually set to a small positive number (such as between 0.01 and 0.1) to ensure a smooth increase in edge strength. The formula for updating the strength value of associated edges is: Ensure that the strength value does not exceed the preset upper limit (usually 1).

[0086] If an ad is deemed invalid, and the graph structured state vector contains explicit negative user feedback information (e.g., the cursor position when the ad is closed, or the user's selected "not interested" reason tag), the system performs a more refined edge weakening operation. Based on the negative user feedback information recorded in the graph structured state vector, negative feedback regions are determined. For example, if a user frequently closes an ad near a specific visual element area (such as an image of a spokesperson or a certain background color), that area is marked. Subsequently, the strength of specific associated edges between creative gene vector nodes associated with the negative feedback region is weakened. The degree of weakening is determined accordingly. A more aggressive approach could be taken to accelerate the elimination of unfavorable portfolio patterns and reduce the magnitude of the impact. The formula is: ,in This is a mitigation factor, ranging from 0.1 to 0.3. The updated formula is: , This represents the edge strength value before the weakening. This approach avoids indiscriminately weakening all edges when there is invalid delivery, and instead applies the penalty precisely to the elements directly related to the negative feedback.

[0087] This application enables dynamic creative feature maps to go beyond simple co-occurrence statistics, improving the intrinsic quality of subsequent candidate creative combinations generated by the system. It guides the online decision-making module to be more inclined to select creative combinations with strong internal synergy and avoid elements that are known to be rejected by users, thereby reducing the probability of ineffective delivery at the source and achieving continuous optimization of creative combination strategies.

[0088] This application further proposes to generate creative selection instructions through a preset strategy function; wherein, the creative selection instructions are creative gene vector recombination instructions, and the specific implementation steps of the generation logic are as follows:

[0089] After the online decision-making module completes the evaluation of a batch of candidate creative combinations, the preset strategy function performs a diagnostic round before executing the regular selection. This identifies candidate creative combinations whose current expected performance value is lower than a preset score threshold. (Preset score threshold) Based on the distribution of historical evaluation scores, for example, the median or 40th percentile of the expected performance values ​​of all historical candidate combinations is taken, with a typical normalization range between 0.4 and 0.6.

[0090] The system calculates the contribution (or sensitivity) of each creative gene vector node within the combination to the expected performance value of the combination. The node with the lowest contribution is identified as the node to be replaced. From the dynamic creative feature map, at least one alternative node with the highest correlation edge strength to the node to be replaced and a higher node weight is searched. The search criteria are based on two points: first, the correlation edge strength between the alternative node and the original low-contribution node must be the highest, ensuring that the alternative element is highly related to the original element in creative semantics or visual style, and the replacement will not cause a break in the combination logic; second, the node weight of the alternative node must be higher, ensuring that the replacement evolves in the direction of better historical performance. The system selects at least one alternative node that meets the above conditions and has the highest weight.

[0091] The system generates a reorganization instruction to replace the original low-contribution node with a replacement node. Subsequently, the reorganization instruction and the resulting new candidate idea combination are returned to the online decision-making module for a new round of performance evaluation. Upon receiving this new combination, the online decision-making module immediately performs a new round of performance evaluation based on the current graph structured state vector, calculating its new expected performance value and decision uncertainty value. This newly generated combination will participate in the subsequent decision-making process of the current time slot (e.g., it may be directly selected due to its improved performance, or it may enter the next round of evaluation).

[0092] By employing a reorganization instruction mechanism, this application achieves a leap from passive selection to proactive construction. When faced with initially underperforming creative solutions, the system does not simply abandon them or passively wait for future exploration. Instead, it proactively initiates a precise and directed improvement attempt based on rich graph knowledge (association strength and node weights). This mechanism enhances the flexibility and intelligence of creative combination generation, improves the efficiency of creative resource utilization, and raises the output quality of the decision-making process.

[0093] This application further proposes that the system also includes a policy embedding module, used for:

[0094] The decision data output by the online decision-making module is monitored. This data includes at least the creative selection instruction and its corresponding decision certainty value. When the graph-structured state vector for the same feature pattern is detected, and the same creative selection instruction is output in a predetermined number of consecutive decisions, with the corresponding decision certainty value consistently exceeding a preset confidence threshold, the monitoring process is successful. At this point, it is determined that the decision-making strategy for that feature pattern has stabilized. Here, "same feature pattern" does not refer to completely identical state vectors, but rather to state vectors falling into a specific cluster or region within a feature space. For example, the system can maintain a dynamic clustering model (such as using incremental K-Means or DBSCAN algorithms) to perform online clustering of continuously generated graph-structured state vectors. When a new state vector arrives, it is assigned to a specific cluster ID, which represents a type of feature pattern.

[0095] The module maintains a decision history record for each identified feature pattern (cluster). When it detects that for the same cluster ID, in a preset number of consecutive decisions (e.g., N=20), the online decision module outputs the same creative selection instruction, and the decision certainty value corresponding to each decision is higher than a preset confidence threshold, the module ensures that the decision is correct. (For example, When the threshold (which can be set according to the business's requirements for decision reliability, typically between 0.85 and 0.95) is reached, the module determines that the decision strategy for that feature pattern has reached a stable state.

[0096] Once stability is determined, the module generates and stores fixed decision rules. These fixed decision rules essentially establish a direct mapping from feature patterns (described by cluster center vectors or cluster boundaries) to creative selection instructions. The rules are stored in a high-speed query rule base. When a new graph-structured state vector is generated, the system first quickly matches it with the feature patterns in the rule base (e.g., calculating the cosine similarity with each cluster center; a match is considered successful if the similarity exceeds a threshold such as 0.95). If a match is successful, the system directly calls the corresponding fixed decision rule to generate creative selection instructions, bypassing the complete model calculation process of the online decision module, thus significantly shortening the decision-making process.

[0097] The introduction of the strategy solidification module enables the system to transform well-validated, high-performance online learning results into efficient execution rules. This reduces the computational load and decision latency when dealing with familiar scenarios, improving overall throughput and response speed. Solidified rules, as carriers of stable knowledge, prevent decision-making fluctuations that may occur in online models due to short-term data volatility, enhancing the certainty and consistency of system decisions in mature scenarios. This allows complex computing resources to be focused on handling novel and uncertain decision-making scenarios, thereby optimizing the system's overall performance and energy efficiency.

[0098] This application further proposes that the system also includes a constraint generation module, used for:

[0099] After the online decision-making module generates an initial set of candidate creative combinations, a graph structure homogeneity detection is performed on the sampled initial candidate creative combinations. The core of the detection is to calculate the graph structure similarity between any two creative combinations. The structure of a creative combination can be characterized by the set of nodes it contains and the set of edges connecting the nodes.

[0100] If a homogeneous subset of graph structures has a similarity higher than a preset similarity threshold, then a representative subset is retained based on the average node weight. Here, similarity... It can be calculated by combining the Jaccard coefficient (which measures the overlap between sets of nodes) and the similarity between sets of edges. ,in, These are the Jaccard coefficients, where N and E represent the node set and edge set, respectively. These are the weighting coefficients. ∈[0,1], for example, set to 0.6. The system sets a preset similarity threshold. (e.g., 0.75), preset similarity threshold The value typically ranges from 0.7 to 0.85 and can be adjusted according to the desired level of diversity.

[0101] For each identified homogeneous subset, the module performs a redundancy removal operation: based on the average node weights, a representative subset is retained. Specifically, the average weight of the nodes in each subset is calculated, and the subset with the highest average weight is retained as the representative, while the remaining subsets are temporarily removed. The purpose of this is to prioritize retaining historically best-performing individuals from the homogeneous options.

[0102] Next, based on the dynamic creative feature map, new alternative creative combinations are generated through resampling, ensuring that the structural differences between the new combination and all currently retained combinations meet a preset diversity requirement. The goal of generating new alternative creative combinations is to ensure sufficient structural differences between them and all currently retained combinations. This difference is assessed by calculating the structural similarity between the new combination and each existing retained combination, and taking the maximum similarity. The system sets a preset diversity requirement: the maximum similarity between the new combination and all existing combinations must be below a diversity threshold. (e.g., 0.4). The module generates combinations that meet this requirement through an iterative sampling process: sampling nodes from the graph (nodes that are intentionally far away from existing retained combinations in the graph can be selected), constructing combinations, and checking whether they meet the diversity threshold. The requirement is to continue until a sufficient number of new alternative combinations are generated.

[0103] Finally, the module merges the representative combinations selected after redundancy removal with the new alternative creative combinations generated by resampling to form a set of candidate creative combinations after redundancy removal. This set is then input into the online decision-making module for subsequent performance and uncertainty evaluation.

[0104] By introducing a constraint generation module, this application avoids wasting limited computational and evaluation resources on evaluating highly similar creative combinations. It improves the efficiency of exploring a broad creative space by not only exploring strategies (through uncertainty) but also actively managing diversity at the source of candidate set generation. This makes the exploration more comprehensive and systematic, helping to discover more potential high-value creative types in long-term operation and reducing the risk of getting trapped in local optima due to sampling bias.

[0105] The following is a specific example of an advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion:

[0106] During a major promotional period, an e-commerce platform launched advertising campaigns for a new athletic shoe product, utilizing this system for dynamic creative evaluation and decision-making. The system is deployed on an Alibaba Cloud GPU-accelerated computing cluster, with dynamic creative feature maps stored in the Neo4j graph database. Real-time data is transmitted via a Kafka message queue (10 broker nodes).

[0107] The state awareness and graphing module first acquires real-time user interaction data: It collects mouse click coordinates and page dwell time on the advertising canvas using a front-end tracking SDK (sampling frequency 50Hz), and parses the gaze dwell heatmap data (two-dimensional matrix resolution 1920×1080). The system uses the Otsu threshold segmentation algorithm to extract region coordinate sequences (e.g., (0.3, 0.4), (0.5, 0.6)) with intensity exceeding a preset threshold (0.6), and performs IoU matching (threshold 0.5) with creative gene vectors representing visual elements in the dynamic creative feature graph (e.g., features extracted from the fc7 layer of the VGG16 network). After successful matching, temporary enhancement weight coefficients are generated. in, =0.2 (the proportion of the high-attention area to the total canvas area). (Average length of stay) Seconds (base duration), α=0.3 (adjustment factor).

[0108] Simultaneously, the system extracts keywords "618 promotion" and "sneaker discount" from external environmental event data (news aggregation API), performs cosine similarity matching (threshold 0.7) with the text creative gene vector in the graph, and generates an impact factor: in, =0.85 (match score) =1200 (Event popularity) =1000 (baseline popularity), β=0.6 (weighting coefficient).

[0109] The system will and Encoded as a graph-structured state vector (128 dimensions).

[0110] The online decision-making module calls the contextual multi-armed gambling machine model (3-layer fully connected network, 1000 training epochs, Adam optimizer) to sample candidate creative combinations (arms) from the graph, using the state vector as input. The model calculates the expected performance value and decision uncertainty value of each arm: when the uncertainty value (0.65) is higher than the first preset threshold (0.5), an exploration strategy is adopted to prioritize creative combinations with an average node weight (0.4) lower than the preset weight threshold (0.5) and an average edge strength (0.8) higher than the preset association strength threshold (0.7) (such as "red shoe body + 'limited-time discount' copy").

[0111] The execution feedback module sends creative selection instructions to the ad execution terminal and receives actual performance data (such as click-through rate and conversion cost) after the campaign period ends. The collaborative evolution update module generates weight adjustment coefficients based on the actual performance data (click-through rate 0.8%, baseline click-through rate 0.6%). The node weights are updated using a time-decay weighted update algorithm: ,in, =0.4 (original weight) (Time decay function, λ=0.1) =0.1 (learning rate) =0.09 (adaptive factor). Simultaneously, the strength of the inter-node connections within this creative combination is enhanced (original strength 0.8, enhancement coefficient 0.05): .

[0112] The constraint generation module performs homogenization detection on the initial candidate creative combinations (similarity threshold 0.75), retains the combination with the highest average node weight, and resamples to generate new combinations with a difference that meets the diversity threshold (0.4) (such as "blue shoe body + 'celebrity style' copywriting"). The strategy solidification module detects that the decision for the feature pattern of "young users + promotional scenario" outputs the same instruction 20 times consecutively, and the deterministic value (0.92) is higher than the preset confidence threshold (0.9), and generates a solidified decision rule: when the state vector matches this feature pattern, directly select the combination of "red shoe body + 'limited-time discount' copywriting" to bypass the model calculation.

[0113] Technical Results: During major promotional periods, the system responds in real-time to changes in user attention (e.g., users pay more attention to the "limited-time discount" copy area), proactively exploring novel creative combinations (e.g., "celebrity endorsement" copy), avoiding the pitfall of "only recommending historically high-click-through-rate creatives" as the local optimum. Through collaborative evolution and updates, the weight of the "limited-time discount" copy node in the dynamic creative feature graph is gradually optimized based on campaign feedback, making the decision model more accurate in predicting the creative preferences of younger user groups. After the strategy is solidified, the decision latency is reduced from 500ms to 100ms, system throughput is increased by 30%, and advertising performance remains stable in complex and dynamic environments, with user click-through rate and conversion rate continuously optimizing over the campaign cycle.

[0114] This embodiment demonstrates the complete application process of the system in e-commerce promotion scenarios. Through multi-source data fusion, dynamic decision-making, and collaborative evolution, it solves the technical problems of traditional advertising systems being unable to respond to market changes in real time and struggling to balance exploration and utilization, and achieves adaptive optimization and continuous evolution of advertising creative placement strategies.

[0115] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.

Claims

1. A dynamic evaluation and intelligent decision-making system for advertising creatives based on multi-source data fusion, characterized by: include: The state awareness and graphing module is used to acquire real-time user interaction behavior data and external environmental event data, and map them to a pre-stored dynamic creative feature graph for matching and mapping processing, outputting a graph-structured state vector. The online decision-making module is used to take the structured state vector of the graph as input, call the pre-trained incremental learning decision model, calculate the expected performance value and decision uncertainty value of multiple candidate creative combinations sampled from the dynamic creative feature graph, and generate creative selection instructions based on the decision uncertainty value through a preset strategy function. The execution feedback module is used to output the creative selection instruction to the ad delivery execution terminal and receive the actual performance data corresponding to the creative selection instruction from the ad delivery execution terminal. The co-evolution update module is used to synchronously update the weights and edge strengths of relevant nodes in the dynamic creative feature graph based on the actual performance data, the corresponding creative selection instructions, and the graph structured state vector, and to update the internal parameters of the incremental learning decision model. The preset strategy function dynamically switches between the exploration strategy and the utilization strategy based on the comparison result between the decision uncertainty value and the preset threshold. The specific steps for generating creative selection instructions through a preset strategy function based on the decision uncertainty value include: When the decision uncertainty value is higher than a first preset threshold, the exploration strategy is adopted; wherein, the exploration strategy includes: calculating the average weight of the creative gene vector nodes contained in each candidate creative combination, and preferentially selecting candidate creative combinations whose average weight is lower than a preset weight threshold and whose average strength of the association edges between nodes in the combination is higher than a preset association strength threshold; When the decision uncertainty value is lower than the second preset threshold, the utilization strategy is adopted to select the candidate creative combination with the highest expected efficiency value; wherein, the second preset threshold is less than the first preset threshold.

2. The advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion according to claim 1, characterized in that: The specific steps for acquiring real-time user interaction data and external environmental event data, and mapping them to a pre-stored dynamic creative feature map for matching and mapping processing include: Obtain the dynamic creative feature map; wherein, the dynamic creative feature map is constructed using creative gene vectors as nodes; The heatmap data of the user's gaze lingering in the advertising canvas area is extracted from the real-time interactive behavior data; Extract the coordinate sequence of regions whose intensity exceeds a preset intensity threshold from the gaze dwell heatmap data; The region coordinate sequence is matched and associated with the creative gene vector representing the visual element; Based on the matching and association results, temporary enhancement weight coefficients for relevant nodes and associated edges are generated, and the enhancement weight coefficients are encoded into the graph structured state vector.

3. The advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion according to claim 2, characterized in that: The specific steps for acquiring real-time user interaction data and external environmental event data, and mapping them to a pre-stored dynamic creative feature map for matching and mapping processing, also include: Extract keywords or sentiment tags from the external environmental event data; Perform semantic similarity matching between the keywords or sentiment tags and creative gene vectors that represent the text's theme or sentiment; Nodes in the dynamic creative feature map whose matching degree exceeds a preset matching threshold are marked as nodes affected by the current external event, and the influence factor of the node is encoded into the structured state vector of the map.

4. The advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion according to claim 1, characterized in that: The incremental learning decision model invoked is a context-based multi-armed gambling machine model, which specifically includes: Each arm of the contextual multi-armed gambling machine model corresponds to a candidate creative combination, which is generated from the dynamic creative feature map based on node weights and associated edge strengths. The context input of the contextual multi-armed gambling machine model is the graph-structured state vector; The prior distribution parameters of the rewards for each arm in the contextual multi-armed gambling machine model are initialized based on the current weight and associated edge strength of the creative gene vector node corresponding to the current arm in the dynamic creative feature map.

5. The advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion according to claim 1, characterized in that: The specific steps for synchronously updating the weights of relevant nodes and the strength of associated edges in the dynamic creative feature graph include: Based on the comparison between the actual performance data and the preset benchmark performance value, a weight adjustment coefficient is generated; Identify the creative gene vector nodes involved in the creative selection instruction; The current weight value of the creative gene vector node is calculated using a time decay-based weighted update algorithm, combined with the weight adjustment coefficient, to obtain the updated node weight value, which is then written back to the dynamic creative feature map.

6. The advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion according to claim 1, characterized in that: The specific steps for synchronously updating the weights of relevant nodes and the strength of associated edges in the dynamic creative feature graph also include: Based on the actual performance data, determine whether the current decision slot is effective. If the result is deemed valid, the strength of all associated edges between the creative gene vector nodes involved in the creative selection instruction is increased. If the result is deemed invalid, the negative feedback region is determined based on the user negative feedback information recorded in the graph structured state vector, thereby weakening the strength of specific association edges between the creative gene vector nodes associated with the negative feedback region.

7. The advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion according to claim 1, characterized in that: Creative selection instructions are generated through a preset strategy function; wherein, the creative selection instructions are creative gene vector recombination instructions, and the generation logic includes: Identify candidate creative combinations whose current expected performance value is lower than a preset score threshold; For the candidate creative combination, identify the creative gene vector node within the combination that contributes the least to the expected performance value; From the dynamic creative feature map, find at least one alternative node that has the highest associated edge strength with the node with the lowest contribution and has a higher node weight; Generate a reorganization instruction to replace the original low-contribution node with the replacement node; The reorganization instruction and the new candidate idea combination formed after reorganization are returned to the online decision-making module for a new round of performance evaluation.

8. The advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion according to claim 1, characterized in that: The system also includes a policy persistence module, used for: Monitor the decision data output by the online decision module, wherein the decision data includes at least the creative selection instruction and the corresponding decision certainty value; When the graph structured state vector for the same feature pattern is detected, and the same creative selection instruction is output in a series of preset number of decisions, and the corresponding decision certainty value is higher than the preset confidence threshold, it is determined that the decision strategy for the feature pattern has been stable. Generate and store solidified decision rules. When the graph structured state vector matches the feature pattern, call the solidified decision rules to generate creative selection instructions and bypass the calculation process of the online decision module. The solidified decision rule establishes a mapping relationship between the feature pattern and the creative selection instruction.

9. The advertising creative dynamic evaluation and intelligent decision-making system based on multi-source data fusion according to claim 1, characterized in that: The system also includes a constraint generation module, used for: The initial candidate creative combinations sampled are subjected to graph structure homogenization detection. If there is a homogenized combination subset with graph structure similarity higher than a preset similarity threshold, a representative combination is retained from the homogenized combination subset based on the average node weight. Based on the dynamic creative feature map, a new alternative creative combination is generated by resampling, which has a structural difference from all currently retained combination maps that meets the preset diversity requirements. The representative combination and the new alternative creative combination are combined to form a redundancy-free candidate creative combination set, which is then input into the online decision-making module.

Images