Regional cultural synergy evaluation method based on artificial intelligence model

Abstract

The present application relates to the field of intelligent management technology, and specifically discloses a method for evaluating regional cultural synergy based on an artificial intelligence model, which uses artificial intelligence technology based on deep learning to perform multimodal data analysis on cultural data related to different regions, captures the multimodal semantic interaction fusion characteristics of regional cultural data, and performs cultural semantic feature interaction analysis on cultural data related to different regions to dig out the cultural differences between different regions, thereby realizing intelligent evaluation of the level of cultural synergy between regions. In this way, the influence of subjective factors can be reduced, the evaluation process can be automated, and a decision-making basis can be provided for regional cultural exchanges and cooperation, so as to better guide and promote the coordinated development of regional culture.

Description

Technical Field

[0001] The present application relates to the field of intelligent management technology, and more specifically, to a method for evaluating regional cultural synergy based on an artificial intelligence model. Background Art

[0002] Regional culture refers to the collection of cultural phenomena such as people's lifestyles, beliefs, arts, customs, etc. in a specific geographical area. This cultural phenomenon not only shapes regional characteristics, but also affects the economic, social and political development of the region to a certain extent. The degree of cultural synergy between regions refers to the coordination and consistency shown by the cultures of different regions in the process of mutual exchange, integration and development. In the context of globalization and regional integration, the degree of cultural synergy between regions directly affects the depth and breadth of regional cooperation, as well as the quality of overall regional development.

[0003] With the advent of globalization and the information age, exchanges and integration between regional cultures are becoming increasingly frequent, and cultural synergy, as a key indicator for measuring the degree of cultural harmony between regions, has received widespread attention. However, at present, the evaluation of regional cultural synergy mainly relies on traditional methods such as questionnaires and expert evaluations. These methods are not only time-consuming and labor-intensive, but also largely affected by subjective factors, resulting in inaccurate and objective evaluation results.

[0004] Therefore, a method for evaluating regional cultural synergy based on an artificial intelligence model is expected. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a method for evaluating regional cultural synergy based on an artificial intelligence model, which uses artificial intelligence technology based on deep learning to perform multimodal data analysis on cultural-related data of different regions, captures the multimodal semantic interaction fusion characteristics of regional cultural-related data, and performs cultural semantic feature interaction analysis on cultural-related data of different regions to explore the cultural differences between different regions, thereby realizing intelligent evaluation of the level of cultural synergy between regions. In this way, the influence of subjective factors can be reduced, the evaluation process can be automated, a decision-making basis can be provided for regional cultural exchanges and cooperation, and regional cultural coordinated development can be better guided and promoted.

[0006] Accordingly, according to one aspect of the present application, a method for evaluating regional cultural synergy based on an artificial intelligence model is provided, which includes:

[0007] Acquire culture-related data of the first region and culture-related data of the second region, wherein the culture-related data includes cultural heritage, artistic expression, lifestyle and social interaction;

[0008] Performing multimodal feature analysis on the culture-related data of the first region and the culture-related data of the second region respectively to obtain a first region text multimodal representation feature matrix and a second region text multimodal representation feature matrix;

[0009] Performing feature interaction analysis on the first region text multimodal representation feature matrix and the second region text multimodal representation feature matrix to obtain a semantic representation feature matrix of inter-regional cultural differences;

[0010] An evaluation result is determined based on the semantic representation feature matrix of the cultural differences between the regions.

[0011] In the above-mentioned regional cultural synergy evaluation method based on the artificial intelligence model, multimodal feature analysis is performed on the culture-related data of the first region and the culture-related data of the second region respectively to obtain a first region text multimodal representation feature matrix and a second region text multimodal representation feature matrix, including: extracting text parts and image parts from the culture-related data of the first region and the culture-related data of the second region to obtain first region culture-related text data, first region culture-related image data, second region culture-related text data and second region culture-related image data; jointly optimizing and encoding the first region culture-related text data and the first region culture-related image data to obtain the first region text multimodal representation feature matrix; jointly optimizing and encoding the second region culture-related text data and the second region culture-related image data to obtain the second region text multimodal representation feature matrix.

[0012] In the above-mentioned regional cultural synergy evaluation method based on artificial intelligence model, the first regional culture-related text data and the first regional culture-related image data are jointly optimized and encoded to obtain the first regional text multimodal representation feature matrix, including: the first regional culture-related text data and the first regional culture-related image data are passed through a first CLIP model including a text encoder and an image encoder to obtain the first regional text multimodal representation feature matrix.

[0013] In the above-mentioned regional cultural synergy evaluation method based on artificial intelligence model, the second region culture-related text data and the second region culture-related image data are jointly optimized and encoded to obtain the second region text multimodal representation feature matrix, including: the second region culture-related text data and the second region culture-related image data are passed through a second CLIP model including a text encoder and an image encoder to obtain the second region text multimodal representation feature matrix.

[0014] In the above-mentioned regional cultural synergy evaluation method based on the artificial intelligence model, feature interaction analysis is performed on the first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix to obtain the inter-regional cultural difference semantic representation feature matrix, including: using the inter-regional cultural feature interaction analysis module to process the first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix to obtain the inter-regional cultural difference semantic representation feature matrix.

[0015] In the above-mentioned regional cultural synergy evaluation method based on the artificial intelligence model, the first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix are processed using the inter-regional cultural feature interaction analysis module to obtain the inter-regional cultural difference semantic representation feature matrix, including: performing matrix expansion on the first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix to obtain the first regional text multimodal representation feature vector and the second regional text multimodal representation feature vector; the first regional text multimodal representation feature vector and the second regional text multimodal representation feature vector are processed using the following feature interaction analysis formula to obtain the inter-regional cultural difference semantic representation feature vector; wherein the feature interaction analysis formula is:

[0016]

[0017] in, is the multimodal characterization feature vector of the first region text, is the multimodal characterization feature vector of the second region text, represents the transpose of the eigenvector, represents the square of the L2 norm of the eigenvector, Indicates the bitwise subtraction of elements. It means the bitwise addition and summation of elements. Representing the semantic representation feature vector of the cultural differences between regions; and reconstructing the dimension of the semantic representation feature vector of the cultural differences between regions to obtain the semantic representation feature matrix of the cultural differences between regions.

[0018] In the above-mentioned regional cultural synergy evaluation method based on the artificial intelligence model, the evaluation result is determined based on the semantic representation feature matrix of the inter-regional cultural differences, including: passing the semantic representation feature matrix of the inter-regional cultural differences through a classifier-based regional synergy evaluator to obtain the evaluation result, and the evaluation result is used to represent the synergy level label.

[0019] Compared with the prior art, the regional cultural synergy evaluation method based on the artificial intelligence model provided by this application uses artificial intelligence technology based on deep learning to perform multimodal data analysis on cultural data related to different regions, captures the multimodal semantic interaction fusion characteristics of regional cultural data, and conducts cultural semantic feature interaction analysis on cultural data related to different regions to dig out the cultural differences between different regions, thereby realizing intelligent evaluation of the level of cultural synergy between regions. In this way, the influence of subjective factors can be reduced, the evaluation process can be automated, and a decision-making basis can be provided for regional cultural exchanges and cooperation, so as to better guide and promote the coordinated development of regional culture. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0021] Figure 1 It is a flowchart of a method for evaluating regional cultural synergy based on an artificial intelligence model according to an embodiment of the present application.

[0022] Figure 2 Schematic diagram of the architecture of the regional cultural synergy evaluation method based on the artificial intelligence model according to an embodiment of the present application.

[0023] Figure 3 This is a flowchart of step S120 in the regional cultural synergy assessment method based on an artificial intelligence model according to an embodiment of the present application.

[0024] Figure 4 This is a flowchart of the training steps in the regional cultural synergy assessment method based on the artificial intelligence model according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] To facilitate the understanding of the embodiments of the present invention, the specific embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the specific embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0026] Figure 1 It is a flowchart of a method for evaluating regional cultural synergy based on an artificial intelligence model according to an embodiment of the present application. Figure 2FIG. 1 is a schematic diagram of the architecture of a method for evaluating regional cultural synergy based on an artificial intelligence model according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to the embodiment of the present application, the method for evaluating regional cultural synergy based on the artificial intelligence model includes the steps of: S110, acquiring culture-related data of a first region and culture-related data of a second region, wherein the culture-related data include cultural heritage, artistic expression, lifestyle and social interaction; S120, performing multimodal feature analysis on the culture-related data of the first region and the culture-related data of the second region respectively to obtain a first region text multimodal representation feature matrix and a second region text multimodal representation feature matrix; S130, performing feature interaction analysis on the first region text multimodal representation feature matrix and the second region text multimodal representation feature matrix to obtain a semantic representation feature matrix of cultural differences between regions; S140, determining an evaluation result based on the semantic representation feature matrix of cultural differences between regions.

[0027] As mentioned in the above background technology, under the current trend of globalization and regional integration, the research and practice of regional cultural synergy evaluation methods are particularly important. However, the existing cultural synergy evaluation methods mainly rely on traditional questionnaires and expert evaluations. These methods are not only time-consuming and labor-intensive, but also easily affected by subjective factors, resulting in inaccurate and objective evaluation results. In response to the above technical problems, the technical concept of this application is to use artificial intelligence technology based on deep learning to perform multimodal data analysis on cultural-related data of different regions, capture the multimodal semantic interaction fusion characteristics of regional cultural-related data, and conduct cultural semantic feature interaction analysis on cultural-related data of different regions to explore the cultural differences between different regions, thereby realizing intelligent evaluation of the level of cultural synergy between regions. In this way, the influence of subjective factors can be reduced, the evaluation process can be automated, and a decision-making basis can be provided for regional cultural exchanges and cooperation, so as to better guide and promote the coordinated development of regional culture.

[0028] In the above-mentioned regional cultural synergy evaluation method based on the artificial intelligence model, the step S110 obtains the cultural-related data of the first region and the cultural-related data of the second region, wherein the cultural-related data includes cultural heritage, artistic expression, lifestyle and social interaction. That is, by obtaining the cultural-related data of different regions for comparative analysis, the cultural differences and similarities between different regions are discovered, and the unique cultural characteristics and common cultural elements of different regions are revealed, so as to realize the evaluation of the cultural synergy between regions. At the same time, considering that culture is a dynamic and multidimensional social phenomenon, it covers many aspects such as history, art, living habits and social interaction. Therefore, in order to more comprehensively understand and compare the cultural characteristics of different regions, in the technical solution of the present application, cultural-related data of cultural heritage, artistic expression, lifestyle and social interaction are obtained. Specifically, cultural heritage is a valuable asset unique to a country or region, representing history, tradition and values. Artistic expression is an important part of culture, reflecting people's aesthetic taste and creativity. Lifestyle is the way and habit of people's daily life, reflecting people's attitude towards life and value orientation under different cultural backgrounds. Social interaction is the way people connect, communicate and interact with each other, reflecting social relations and cultural identity. By acquiring these culturally relevant data, we can gain a comprehensive understanding of the cultural characteristics of different regions, which helps us to gain a deeper understanding and compare the cultural characteristics of different regions, thereby achieving intelligent evaluation and comparative analysis of cultural synergy between regions.

[0029] In the above-mentioned regional cultural synergy evaluation method based on the artificial intelligence model, the step S120 performs multimodal feature analysis on the culture-related data of the first region and the culture-related data of the second region to obtain the first region text multimodal representation feature matrix and the second region text multimodal representation feature matrix. Specifically, Figure 3 FIG. 1 is a flowchart of step S120 in the method for evaluating regional cultural synergy based on an artificial intelligence model according to an embodiment of the present application. Figure 3 As shown, the step S120 includes: S121, extracting text parts and image parts from the culture-related data of the first region and the culture-related data of the second region to obtain first-region culture-related text data, first-region culture-related image data, second-region culture-related text data and second-region culture-related image data; S122, jointly optimizing and encoding the first-region culture-related text data and the first-region culture-related image data to obtain a multimodal representation feature matrix of the first-region text; S123, jointly optimizing and encoding the second-region culture-related text data and the second-region culture-related image data to obtain a multimodal representation feature matrix of the second-region text.

[0030] Specifically, the step S121 extracts the text part and the image part from the culture-related data of the first region and the culture-related data of the second region to obtain the first region culture-related text data, the first region culture-related image data, the second region culture-related text data, and the second region culture-related image data. It should be understood that the culture-related data is a multimodal data, which usually contains data in multiple forms, such as text data and image data. Among them, text data can provide detailed text information, while image data has intuitive and visual characteristics, and can present cultural related information more vividly through image display. However, considering that text and images are different data forms, it is impossible to use a unified data processing method for analysis and processing. Therefore, the text part and the image part are further extracted from the culture-related data of the first region and the culture-related data of the second region, so as to facilitate independent data analysis of the text part and the image part, and then perform information fusion to comprehensively utilize text information and image information, thereby enriching the dimension of the data and improving the information content and expression ability of the data.

[0031] Specifically, in step S122, the first region culture-related text data and the first region culture-related image data are jointly optimized and encoded to obtain a multimodal characterization feature matrix of the first region text. In a specific example of the present application, the encoding method for jointly optimizing the encoding of the first region culture-related text data and the first region culture-related image data is to pass the first region culture-related text data and the first region culture-related image data through a first CLIP model including a text encoder and an image encoder to obtain a multimodal characterization feature matrix of the first region text. It should be understood that the CLIP model is a multimodal model that can process text data and image data simultaneously, which includes a text encoder and an image encoder, and can encode text and images simultaneously. In the technical solution of the present application, a first CLIP model including a text encoder and an image encoder is used to process the first region culture-related text data and the first region culture-related image data. The first CLIP model can semantically encode the first region culture-related text data and image data through the text encoder and the image encoder respectively, capture the image semantic features and the text semantic features, and use the contrastive learning method to learn a shared multimodal embedding space, so that the image semantic features and the text semantic features can be represented and jointly encoded in the same feature space, so as to learn the correspondence between the image semantic features and the text semantic features, and while comprehensively utilizing the semantic features of the image data and the text data, maintain the semantic relevance and consistency between the text data and the image data, so as to more comprehensively characterize the culture-related data of the first region and improve the quality and accuracy of data representation.

[0032] Specifically, the step S123 performs joint optimization encoding on the second region culture-related text data and the second region culture-related image data to obtain the second region text multimodal representation feature matrix. In a specific example of the present application, the encoding method for jointly optimizing the encoding of the second region culture-related text data and the second region culture-related image data is to pass the second region culture-related text data and the second region culture-related image data through a second CLIP model including a text encoder and an image encoder to obtain the second region text multimodal representation feature matrix. That is, similarly, the second region culture-related text data and the second region culture-related image data are jointly encoded using a second CLIP model including a text encoder and an image encoder to capture the semantic feature consistency and correlation between the second region culture-related text data and image data, and integrate the text information and the image information to generate a second region text multimodal representation feature matrix, thereby more comprehensively representing the culture-related data of the second region.

[0033] In the above-mentioned regional cultural synergy evaluation method based on the artificial intelligence model, the step S130 performs feature interaction analysis on the first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix to obtain the inter-regional cultural difference semantic representation feature matrix. In a specific example of the present application, the encoding method for performing feature interaction analysis on the first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix is ​​to use the inter-regional cultural feature interaction analysis module to process the first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix to obtain the inter-regional cultural difference semantic representation feature matrix. It should be understood that the first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix respectively represent the cultural-related data information of the first region and the second region, and by interactively analyzing the text multimodal representation feature matrices of the two regions, the cultural similarities and differences between different regions can be revealed. In the technical solution of the present application, the inter-regional cultural feature interaction analysis module is a generalization of a linear regression model that can model the interaction between features, especially when processing high-dimensional sparse data, it shows extremely high efficiency. Here, the inter-regional cultural feature interaction analysis module is used to process the first region text multimodal representation feature matrix and the second region text multimodal representation feature matrix, which can effectively capture the deep semantic interaction relationship between the cultural-related data information of the first region and the second region, and explore the feature differences between the two, thereby more accurately characterizing the degree of cultural synergy between the first region and the second region.

[0034] Specifically, the step S130 includes: performing matrix expansion on the first region text multimodal representation feature matrix and the second region text multimodal representation feature matrix to obtain the first region text multimodal representation feature vector and the second region text multimodal representation feature vector; processing the first region text multimodal representation feature vector and the second region text multimodal representation feature vector using the following feature interaction analysis formula to obtain a semantic representation feature vector of cultural differences between regions; wherein the feature interaction analysis formula is:

[0035]

[0036] in, is the multimodal characterization feature vector of the first region text, is the multimodal characterization feature vector of the second region text, represents the transpose of the eigenvector, represents the square of the L2 norm of the eigenvector, Indicates the bitwise subtraction of elements. It means the bitwise addition and summation of elements. Representing the semantic representation feature vector of the cultural differences between regions; and reconstructing the dimension of the semantic representation feature vector of the cultural differences between regions to obtain the semantic representation feature matrix of the cultural differences between regions.

[0037] In the above-mentioned regional cultural synergy evaluation method based on the artificial intelligence model, the step S140 determines the evaluation result based on the semantic representation feature matrix of the inter-regional cultural differences. In a specific example of the present application, the implementation method of determining the evaluation result based on the semantic representation feature matrix of the inter-regional cultural differences is to pass the semantic representation feature matrix of the inter-regional cultural differences through a regional synergy evaluator based on a classifier to obtain the evaluation result, and the evaluation result is used to represent the synergy level label. It should be understood that the inter-regional synergy evaluator is based on a classification algorithm and can assign input data to predefined categories or labels. In the technical solution of the present application, the inter-regional cultural difference semantic representation feature matrix contains the cultural difference features between the first region and the second region, which are input into the inter-regional synergy evaluator based on the classifier. The inter-regional synergy evaluator based on the classifier can map the inter-regional cultural difference semantic representation feature matrix to a preset label used to represent the inter-regional cultural synergy level by learning the cultural difference features in the inter-regional cultural difference semantic representation feature matrix, thereby obtaining the cultural synergy evaluation result between the first region and the second region. In this way, cultural differences between regions are labeled and the degree of cultural differences between different regions is quantified, thereby providing a decision-making basis for regional cultural exchanges and cooperation, and better guiding and promoting the coordinated development of regional cultures.

[0038] In the technical solution described above, the first region text multimodal representation feature matrix represents the cross-modal global interactive fusion features of the semantic coding features of the first region culture-related text data and the image semantic coding features of the first region culture-related image data based on the feature value granularity. The second region text multimodal representation feature matrix represents the cross-modal global interactive fusion features of the semantic coding features of the second region culture-related text data and the image semantic coding features of the second region culture-related image data based on the feature value granularity. However, considering that there are modal differences and source data content differences between the first region culture-related text data and the first region culture-related image data, and between the second region culture-related text data and the second region culture-related image data, after text semantic coding, image semantic coding and cross-modal interaction based on the CLIP model, there is a fine-grained misalignment of feature distribution between the second region text multimodal representation feature matrix and the inter-regional cultural difference semantic representation feature matrix.

[0039] In this way, when the inter-regional cultural feature interaction analysis module is used to process the second regional text multimodal representation feature matrix and the first regional text multimodal representation feature matrix, the obtained inter-regional cultural difference semantic representation feature matrix will have more significant local distribution inconsistencies in feature distribution, causing feature distribution regression mode shift, that is, causing the inter-regional cultural difference semantic representation feature matrix to overflow the evaluation result obtained by the classifier-based inter-regional synergy evaluator, affecting its accuracy.

[0040] Based on this, in a preferred embodiment of the present application, the semantic representation feature matrix of inter-regional cultural differences is input into an inter-regional synergy evaluator based on a classifier to obtain an evaluation result, specifically including:

[0041] Determine the semantic representation feature matrix of the cultural differences between regions and the input of the regional synergy evaluator based on the classifier The corresponding coordination level labels label probability values, and calculate the The square root of the sum of the squares of the label probability values ​​is used to obtain the type distribution probability value;

[0042] The inter-regional cultural difference semantic representation feature matrix is ​​passed through a probabilistic activation function to obtain a probabilistic inter-regional cultural difference semantic representation feature matrix;

[0043] Subtracting the probabilistic semantic representation feature matrix of regional cultural differences from the unit feature matrix, and using the feature matrix resulting from the subtraction as a base to calculate a power function with the type distribution probability value as an exponent;

[0044] Performing a point multiplication of the power function feature matrix and the probabilistic feature matrix of semantic representation of cultural differences between regions to obtain an intermediate feature matrix of semantic representation of cultural differences between regions;

[0045] Taking the absolute value of each eigenvalue of the probabilistic semantic representation feature matrix of inter-regional cultural differences as the base, calculating a power function with the inverse of the type distribution probability value as the exponent, and summing all eigenvalues ​​to obtain the biased eigenvalue of the semantic representation of inter-regional cultural differences;

[0046] After multiplying the bias feature value of the semantic representation of cultural differences between regions by a weight as a hyperparameter, the result is interpolated with the intermediate feature matrix of the semantic representation of cultural differences between regions to obtain an optimized feature matrix of the semantic representation of cultural differences between regions;

[0047] The optimized inter-regional cultural difference semantic representation feature matrix is ​​input into the inter-regional synergy evaluator based on the classifier to obtain the evaluation result.

[0048] In which, let the probabilistic semantic representation feature matrix of regional cultural differences be recorded as , then the optimized semantic representation feature matrix of regional cultural differences is It is expressed as:

[0049]

[0050] in, Representing the probabilistic semantic representation feature matrix of regional cultural differences, is the unit eigenvalue matrix with all eigenvalues ​​equal to one, and , is the probabilistic semantic representation feature matrix of regional cultural differences For each eigenvalue of Represents the type distribution probability value, is a power function that calculates the eigenvalues ​​of each position in the eigenvector with the type distribution probability value as an exponent, It means point multiplication by position. It means to make a difference by position. It means adding by position. is the optimized semantic representation feature matrix of inter-regional cultural differences, and, The label probability value is expressed as , then the type distribution probability value It is expressed as:

[0051]

[0052] in, for The type label of the change region corresponds to the The label probability value, Indicates the type distribution probability value.

[0053] That is, for the semantic characterization feature matrix of inter-regional cultural differences, based on the convergence limit of the probability distribution of the Cauchy-Hadamard form power series, the feature set of the semantic characterization feature matrix of inter-regional cultural differences is represented by a formal power series based on the convergence probability, and a set convergence bounded bias represented by the convergence radius based on each eigenvalue of the semantic characterization feature matrix of inter-regional cultural differences is added. Under a unified probability convergence limit standard, this can avoid the invalid overflow of the classification results caused by the deviation of the distribution pattern of the semantic characterization feature matrix of inter-regional cultural differences during the regression process of the inter-regional synergy evaluator based on the classifier due to the gradient propagation of the model parameters based on the convergence probability, thereby improving the accuracy of the evaluation results obtained by the inter-regional synergy evaluator based on the classifier.

[0054] It should be understood that before using the above-mentioned neural network model, it is necessary to train the first CLIP model including the text encoder and the image encoder, the second CLIP model including the text encoder and the image encoder, the inter-regional cultural feature interaction analysis module, and the inter-regional synergy evaluator based on the classifier. That is to say, in the regional cultural synergy evaluation method based on the artificial intelligence model of the present application, it also includes a training step of training the first CLIP model including the text encoder and the image encoder, the second CLIP model including the text encoder and the image encoder, the inter-regional cultural feature interaction analysis module, and the inter-regional synergy evaluator based on the classifier.

[0055] Figure 4 Flowchart of the training steps in the method for evaluating regional cultural synergy based on an artificial intelligence model according to an embodiment of the present application. Figure 4As shown, the training step includes: S210, obtaining training data, the training data including training culture-related data of the first region and training culture-related data of the second region, and the true value of the evaluation result, wherein the training culture-related data includes cultural heritage, artistic expression, lifestyle and social interaction; S220, extracting text parts and image parts from the training culture-related data of the first region and the training culture-related data of the second region to obtain training first region culture-related text data, training first region culture-related image data, training second region culture-related text data and training second region culture-related image data; S230, passing the training first region culture-related text data and the training first region culture-related image data through the first CLIP model including a text encoder and an image encoder to obtain a training first region text multimodal representation feature matrix; S240, passing the training second region culture-related text data and the training second region culture-related image data through the first CLIP model including a text encoder and an image encoder encoder and the second CLIP model of the image encoder to obtain the training second region text multimodal representation feature matrix; S250, use the inter-regional cultural feature interaction analysis module to process the training first region text multimodal representation feature matrix and the training second region text multimodal representation feature matrix to obtain the training inter-regional cultural difference semantic representation feature matrix; S260, pass the training inter-regional cultural difference semantic representation feature matrix through the classifier-based inter-regional synergy evaluator to obtain the classification loss function value; S270, use the classification loss function value to train the first CLIP model including the text encoder and the image encoder, the second CLIP model including the text encoder and the image encoder, the inter-regional cultural feature interaction analysis module and the classifier-based inter-regional synergy evaluator, wherein in each round of the training iteration, the training inter-regional cultural difference semantic representation feature vector obtained by expanding the training inter-regional cultural difference semantic representation feature matrix is ​​iteratively optimized.

[0056] In the technical solution described above, the multimodal representation feature matrix of the training first region text represents the cross-modal global interactive fusion features of the semantic coding features of the training first region culture-related text data and the image semantic coding features of the training first region culture-related image data based on the eigenvalue granularity. The multimodal representation feature matrix of the training second region text represents the cross-modal global interactive fusion features of the semantic coding features of the training second region culture-related text data and the image semantic coding features of the training second region culture-related image data based on the eigenvalue granularity. However, considering that there are modal differences and source data content differences between the training first region culture-related text data and the training first region culture-related image data, and between the training second region culture-related text data and the training second region culture-related image data, after text semantic coding, image semantic coding and cross-modal interaction based on the CLIP model, there is a fine-grained misalignment of feature distribution between the training second region text multimodal representation feature matrix and the training first region text multimodal representation feature matrix.

[0057] In this way, when the inter-regional cultural feature interaction analysis module is used to process the training second region text multimodal representation feature matrix and the training first region text multimodal representation feature matrix, the obtained training inter-regional cultural difference semantic representation feature matrix will have more significant inconsistencies between local distributions of feature distributions, thereby affecting the convergence effect of the training inter-regional cultural difference semantic representation feature matrix to the class probability density space when it is classified by the classifier-based inter-regional synergy evaluator.

[0058] Therefore, in the technical solution of the present application, each time the semantic representation feature matrix of cultural differences between training regions is classified and iterated by the inter-regional synergy evaluator based on the classifier, the semantic representation feature vector of cultural differences between training regions obtained by expanding the semantic representation feature matrix of cultural differences between training regions is optimized by the following optimization formula, wherein the optimization formula is:

[0059]

[0060]

[0061] in, is the semantic representation feature vector of cultural differences between the training regions, and They are the semantic representation feature vectors of cultural differences between the training regions No. and The eigenvalues ​​of the positions, is the mean between every two eigenvalues ​​of the semantic representation feature vector of cultural differences between the training regions, is the mean matrix of the semantic representation feature vectors of cultural differences between the training regions, is the average difference between every two eigenvalues ​​of the semantic representation feature vector of cultural differences between the training regions, is the mean difference matrix of the semantic representation feature vectors of cultural differences between the training regions, It is the optimized semantic representation feature vector of cultural differences between training regions.

[0062] That is, by introducing the semantic representation feature vector of cultural differences between the training regions The local statistical information distribution is used as an external information source to enhance the retrieval of the feature vector, so as to avoid the local overflow information distribution leading to the cultural difference semantic representation feature vector between the training regions based on the local statistical dense information structure. distribution hallucination, thereby obtaining the semantic representation feature vector of cultural differences between the training regions Information credible response reasoning based on local distribution group dimension preservation to obtain the semantic representation feature vector of cultural differences between the training regions Based on the credible distribution response in the probability density space of discretized local feature distribution, the convergence effect of the probability density space is improved to improve the training speed and the accuracy of the training results.

[0063] In summary, the regional cultural synergy evaluation method based on the artificial intelligence model according to the embodiment of the present application is explained, which uses artificial intelligence technology based on deep learning to perform multimodal data analysis on cultural data related to different regions, captures the multimodal semantic interaction fusion characteristics of regional cultural data, and conducts cultural semantic feature interaction analysis on cultural data related to different regions to explore the cultural differences between different regions, thereby realizing intelligent evaluation of the level of cultural synergy between regions. In this way, the influence of subjective factors can be reduced, the evaluation process can be automated, and a decision-making basis can be provided for regional cultural exchanges and cooperation, so as to better guide and promote the coordinated development of regional culture.

[0064] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0065] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the module division is only a logical function division, and there may be other division methods in actual implementation. The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0066] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0067] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0068] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for evaluating regional cultural synergy based on an artificial intelligence model, characterized in that: include: Acquire cultural data related to the first region and cultural data related to the second region, wherein the cultural data include cultural heritage, artistic expression, lifestyle and social interaction, and the synergy assessment includes obtaining cultural data related to different regions and conducting comparative analysis to discover cultural differences and similarities between different regions, reveal unique cultural characteristics and common cultural elements of different regions, and thus achieve an assessment of cultural synergy between regions, wherein the cultural heritage includes history, tradition and values, the artistic expression includes aesthetic taste and creativity, the lifestyle includes people's daily lifestyle and habits, and the social interaction includes the way people connect, communicate and interact with each other; Performing multimodal feature analysis on the culture-related data of the first region and the culture-related data of the second region respectively to obtain a first region text multimodal representation feature matrix and a second region text multimodal representation feature matrix; Performing feature interaction analysis on the first region text multimodal representation feature matrix and the second region text multimodal representation feature matrix to obtain a semantic representation feature matrix of inter-regional cultural differences; Determining an evaluation result based on the semantic representation feature matrix of the cultural differences between the regions; Based on the semantic representation feature matrix of the cultural differences between the regions, the evaluation results are determined, including: The semantic representation feature matrix of inter-regional cultural differences is passed through an inter-regional synergy evaluator based on a classifier to obtain the evaluation result, and the evaluation result is used to represent a synergy level label; The inter-regional cultural difference semantic representation feature matrix is ​​input into the inter-regional coordination degree evaluator based on the classifier to obtain the evaluation result, specifically including: Determine the semantic representation feature matrix of the cultural differences between regions and the input of the regional synergy evaluator based on the classifier The corresponding coordination level labels label probability values, and calculate the The square root of the sum of the squares of the label probability values ​​is used to obtain the type distribution probability value; The inter-regional cultural difference semantic representation feature matrix is ​​passed through a probabilistic activation function to obtain a probabilistic inter-regional cultural difference semantic representation feature matrix; Subtracting the probabilistic semantic representation feature matrix of regional cultural differences from the unit feature matrix, and using the feature matrix resulting from the subtraction as a base to calculate a power function with the type distribution probability value as an exponent; Performing a point multiplication of the power function feature matrix and the probabilistic feature matrix of semantic representation of cultural differences between regions to obtain an intermediate feature matrix of semantic representation of cultural differences between regions; Taking the absolute value of each eigenvalue of the probabilistic semantic representation feature matrix of inter-regional cultural differences as the base, calculating a power function with the inverse of the type distribution probability value as the exponent, and summing the calculation results to obtain the biased eigenvalue of the semantic representation of inter-regional cultural differences; After multiplying the bias feature value of the semantic representation of cultural differences between regions by a weight as a hyperparameter, the result is interpolated with the intermediate feature matrix of the semantic representation of cultural differences between regions to obtain an optimized feature matrix of the semantic representation of cultural differences between regions; Inputting the optimized inter-regional cultural difference semantic representation feature matrix into the inter-regional synergy evaluator based on the classifier to obtain an evaluation result; Let the probabilistic semantic representation feature matrix of regional cultural differences be recorded as , then the optimized semantic representation feature matrix of regional cultural differences is It is expressed as: in, Representing the probabilistic semantic representation feature matrix of regional cultural differences, is the unit eigenvalue matrix with all eigenvalues ​​equal to one, and , is the probabilistic semantic representation feature matrix of regional cultural differences For each eigenvalue of Represents the type distribution probability value, is a power function that calculates the eigenvalues ​​of each position in the eigenvector with the type distribution probability value as an exponent, It means point multiplication by position. It means to make a difference by position. It means adding by position. is the optimized semantic representation feature matrix of inter-regional cultural differences, and, The label probability value is expressed as , then the type distribution probability value It is expressed as: in, for The type label of the change region corresponds to the The label probability value, Indicates the type distribution probability value; Performing multimodal feature analysis on the culture-related data of the first region and the culture-related data of the second region respectively to obtain a first region text multimodal representation feature matrix and a second region text multimodal representation feature matrix, including: Extracting a text portion and an image portion from the culture-related data of the first region and the culture-related data of the second region to obtain first-region culture-related text data, first-region culture-related image data, second-region culture-related text data, and second-region culture-related image data; Performing joint optimization encoding on the first region culture-related text data and the first region culture-related image data to obtain a multimodal representation feature matrix of the first region text; The second region culture-related text data and the second region culture-related image data are jointly optimized and encoded to obtain a second region text multimodal representation feature matrix.

2. The regional cultural synergy evaluation method based on artificial intelligence model according to claim 1 is characterized in that: The first region culture-related text data and the first region culture-related image data are jointly optimized and encoded to obtain a first region text multimodal representation feature matrix, including: The first region culture-related text data and the first region culture-related image data are passed through a first CLIP model including a text encoder and an image encoder to obtain a first region text multimodal representation feature matrix.

3. The regional cultural synergy evaluation method based on artificial intelligence model according to claim 2 is characterized in that: The second region culture-related text data and the second region culture-related image data are jointly optimized and encoded to obtain a second region text multimodal representation feature matrix, including: The second region culture-related text data and the second region culture-related image data are passed through a second CLIP model including a text encoder and an image encoder to obtain a second region text multimodal representation feature matrix.

4. The regional cultural synergy evaluation method based on artificial intelligence model according to claim 3 is characterized in that: Performing feature interaction analysis on the first region text multimodal representation feature matrix and the second region text multimodal representation feature matrix to obtain a semantic representation feature matrix of inter-regional cultural differences, including: The first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix are processed using an inter-regional cultural feature interaction analysis module to obtain the inter-regional cultural difference semantic representation feature matrix.

5. The regional cultural synergy evaluation method based on artificial intelligence model according to claim 4 is characterized in that: Using the inter-regional cultural feature interaction analysis module to process the first regional text multimodal representation feature matrix and the second regional text multimodal representation feature matrix to obtain the inter-regional cultural difference semantic representation feature matrix, including: Performing matrix expansion on the first region text multimodal representation feature matrix and the second region text multimodal representation feature matrix to obtain a first region text multimodal representation feature vector and a second region text multimodal representation feature vector; The first region text multimodal representation feature vector and the second region text multimodal representation feature vector are processed using the following feature interaction analysis formula to obtain a semantic representation feature vector of cultural differences between regions; wherein the feature interaction analysis formula is: in, is the multimodal characterization feature vector of the first region text, is the multimodal characterization feature vector of the second region text, represents the transpose of the eigenvector, represents the square of the L2 norm of the eigenvector, Indicates the bitwise subtraction of elements. It means the bitwise addition and summation of elements. A feature vector representing the semantic representation of the cultural differences between the regions; The dimension of the semantic representation feature vector of the cultural differences between regions is reconstructed to obtain the semantic representation feature matrix of the cultural differences between regions.

Images