CN117576504A - Training method for social media fake news detection model - Google Patents

Abstract

The invention provides a training method of a social media false news detection model, which comprises the following steps: acquiring a plurality of news propagation graphs of a first field in a training set; the news propagation graph is used for representing a news propagation path; acquiring a plurality of news propagation graphs of a second field in the training set; the ratio of the number of news maps of the second domain to the number of news maps of the first domain is less than a threshold; training the social media false news detection model according to the plurality of news propagation graphs in the first field, the plurality of news propagation graphs in the second field and the first target loss function to obtain a social media false news detection model trained based on the training set; the social media false news detection model is used for detecting the authenticity of news; the first objective loss function is determined by the classification loss, the global contrast loss, and the local contrast loss. The method of the invention realizes the accuracy of false news detection in the second field with less data quantity and shorter propagation time.

Description

Training methods for social media fake news detection models

Technical Field

The present invention relates to the field of data processing technology, and in particular to the training of a social media fake news detection model.

Background Art

With the rapid development of the Internet, social media has become an important platform for people to obtain information, express opinions, and communicate in daily life. However, with the growth of the number of users on social platforms, if a piece of fake news becomes a hot topic, it will affect social stability and bring potential economic loss risks after being discussed and spread by a large number of users. In order to determine the authenticity of news, people have proposed the task of social media fake news detection, which aims to determine the authenticity of news in social media.

In the related art, the automatic fake news detection model trained in the high-resource domain with a large amount of data can accurately detect the high-resource domain; however, for emerging fields where emergencies occur, due to insufficient data, the accuracy of automatic fake news detection is low. Therefore, how to accurately detect the authenticity of news in the low-resource domain when there is sufficient high-resource data and less low-resource data is a problem that technicians in this field need to solve urgently.

Summary of the invention

In view of the problems in the prior art, an embodiment of the present invention provides a training method for a social media fake news detection model.

Specifically, the embodiment of the present invention provides the following technical solutions:

In a first aspect, an embodiment of the present invention provides a method for training a social media fake news detection model, comprising:

Obtain multiple news diffusion graphs in the first field of the training set; the news diffusion graph is used to represent the diffusion path of news;

Acquire multiple news diffusion graphs in the second field in the training set; the ratio of the number of news diffusion graphs in the second field to the number of news diffusion graphs in the first field is less than a threshold;

According to multiple news propagation graphs in the first field, multiple news propagation graphs in the second field and the first target loss function, a social media fake news detection model is trained to obtain a social media fake news detection model trained based on the training set; the social media fake news detection model is used to detect the authenticity of news; the first target loss function is determined by classification loss, global contrast loss and local contrast loss; the global contrast loss represents the degree of correlation between node features in the news propagation graph, node features in the first type of augmented graph of the news propagation graph and news propagation graph features; the local contrast loss represents the degree of correlation between node features in the second type of augmented graph of the news propagation graph and node features in the third type of augmented graph of the news propagation graph; the classification loss represents the accuracy of the classification result.

Furthermore, the social media fake news detection model includes at least one of the following:

Feature extraction module: The feature extraction module is used to extract the features of the news dissemination graph;

Classification module: The classification module is used to predict the authenticity of social media news corresponding to the news diffusion graph based on the characteristics of the news diffusion graph;

Self-supervised learning module; the self-supervised learning module is used to determine the global contrast loss based on the node features in the news diffusion graph, the node features in the first type of augmented graph of the news diffusion graph, and the news diffusion graph features; and determine the local contrast loss based on the node features in the second type of augmented graph of the news diffusion graph and the node features in the third type of augmented graph of the news diffusion graph.

Furthermore, the social media fake news detection model is trained based on the following method:

Input multiple news propagation graphs in the first domain and multiple news propagation graphs in the second domain in the training set into the social media fake news detection model, and output the authenticity detection results of the social media news corresponding to the news propagation graphs; obtain the classification loss of the social media fake news detection model based on the authenticity detection results of the social media news and the label information of the news; the label information is used to mark the authenticity of the news;

Input multiple news propagation graphs in the first field in the training set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the first field, the second type of augmented graphs corresponding to the multiple news propagation graphs in the first field, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the first field into the social media fake news detection model to obtain the global contrast loss and local contrast loss of the multiple news in the first field;

Input multiple news propagation graphs in the second field in the training set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the second field, the second type of augmented graphs corresponding to the multiple news propagation graphs in the second field, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the second field into the social media fake news detection model to obtain the global contrast loss and local contrast loss of the multiple news in the second field;

The weighted sum of the classification loss of the social media fake news detection model, the global contrast loss and local contrast loss of multiple news in the first field, and the global contrast loss and local contrast loss of multiple news in the second field is taken as the value of the first objective loss function;

Based on the value of the first objective loss function, the social media fake news detection model is trained to obtain a social media fake news detection model trained based on the training set.

Furthermore, the social media fake news detection model also includes:

Data adaptive constraint module: The data adaptive constraint module is used to determine the difference between the news diffusion graph features in the training set and the news diffusion graph features in the test set.

Furthermore, after training the social media fake news detection model based on the value of the first objective loss function to obtain the social media fake news detection model trained based on the training set, the method further includes:

Input multiple news propagation graphs in the second field in the test set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set, the second type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set into the social media fake news detection model trained based on the training set, and obtain the global contrast loss and local contrast loss of the multiple news in the second field in the test set;

According to the difference between the news propagation graph features of the second field in the training set and the news propagation graph features of the second field in the test set, the weighted sum of the global contrast loss and the local contrast loss of multiple news in the second field in the test set is used as the value of the second objective loss function;

Based on the value of the second objective loss function, the social media fake news detection model is trained to obtain a social media fake news detection model based on the potential features of the test set.

In a second aspect, an embodiment of the present invention further provides a method for detecting fake news on social media, comprising:

Obtain a news dissemination graph of a second field to be detected;

The news propagation graph of the second field to be detected is input into the social media fake news detection model to obtain the authenticity detection result of the social media news corresponding to the news propagation graph of the second field; the social media fake news detection model is trained based on the training method of the social media fake news detection model of the first aspect.

In a third aspect, an embodiment of the present invention further provides a social media fake news detection device, comprising:

Obtain a news dissemination graph of a second field to be detected;

The news propagation graph of the second field to be detected is input into the social media fake news detection model to obtain the authenticity detection result of the social media news corresponding to the news propagation graph of the second field; the social media fake news detection model is trained based on the training method of the social media fake news detection model as described in the first aspect.

In a fourth aspect, an embodiment of the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the training method for the social media fake news detection model as described in the first aspect or the social media fake news detection method as described in the second aspect is implemented.

In a fifth aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the training method for a social media fake news detection model as described in the first aspect or the social media fake news detection method as described in the second aspect.

In a sixth aspect, an embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the training method for the social media fake news detection model as described in the first aspect or the social media fake news detection method as described in the second aspect.

The training method of the social media fake news detection model provided by the embodiment of the present invention trains the social media fake news detection model according to multiple news propagation graphs in the first field, multiple news propagation graphs in the second field and the first target loss function, so that the trained social media fake news detection model can not only accurately extract the feature information of the news in the first field and the global contrast loss between the nodes in the news propagation graph and the news propagation graph, and the local contrast loss between the augmented graph nodes in the news propagation graph, but also accurately extract the feature information of the news in the second field and the global contrast loss between the nodes in the news propagation graph and the news propagation graph, and the local contrast loss between the augmented graph nodes in the news propagation graph, so that the trained social media fake news detection model can accurately and comprehensively extract the feature information of the fake news in the second field with less data volume and shorter propagation time, effectively promote and improve the detection accuracy of fake news, so that the trained social media fake news detection model can accurately detect the authenticity of the news in the second field with less data volume and shorter propagation time, and improve the accuracy of fake news detection in the second field with less data volume and shorter propagation time.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

FIG1 is a flow chart of a method for training a social media fake news detection model according to an embodiment of the present invention;

FIG2 is a second flow chart of a method for training a social media fake news detection model provided by an embodiment of the present invention;

FIG3 is a third flow chart of a method for training a social media fake news detection model provided by an embodiment of the present invention;

FIG4 is a fourth flow chart of a method for training a social media fake news detection model provided by an embodiment of the present invention;

FIG5 is a schematic diagram of the structure of a training device for a social media fake news detection model provided by an embodiment of the present invention;

FIG. 6 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention.

DETAILED DESCRIPTION

In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

The method of the embodiment of the present invention can be applied to fake news detection scenarios, thereby improving the accuracy of fake news detection in the second field with a small amount of data and a short propagation time.

In the related art, the automatic fake news detection model trained in the high-resource domain with a large amount of data can accurately detect the high-resource domain; however, for emerging fields where emergencies occur, due to insufficient data, the accuracy of automatic fake news detection is low. Therefore, how to accurately detect the authenticity of news in the low-resource domain when there is sufficient high-resource data and less low-resource data is a problem that technicians in this field need to solve urgently.

The training method of the social media fake news detection model of the embodiment of the present invention trains the social media fake news detection model according to multiple news propagation graphs in the first field, multiple news propagation graphs in the second field and the first target loss function, so that the trained social media fake news detection model can not only accurately extract the feature information of the news in the first field and the global contrast loss between the nodes in the news propagation graph and the news propagation graph, and the local contrast loss between the augmented graph nodes in the news propagation graph, but also accurately extract the feature information of the news in the second field and the global contrast loss between the nodes in the news propagation graph and the news propagation graph, and the local contrast loss between the augmented graph nodes in the news propagation graph, so that the trained social media fake news detection model can accurately and comprehensively extract the feature information of the fake news in the second field with less data volume and shorter propagation time, effectively promoting and improving the detection accuracy of fake news, so that the trained social media fake news detection model can accurately detect the authenticity of the news in the second field with less data volume and shorter propagation time, thereby improving the accuracy of fake news detection in the second field with less data volume and shorter propagation time.

The technical solution of the present invention is described in detail with reference to specific embodiments in conjunction with Figures 1 to 6. The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

FIG1 is a flow chart of an embodiment of a method for training a social media fake news detection model provided by an embodiment of the present invention. As shown in FIG1 , the method provided by this embodiment includes:

Step 101, obtaining multiple news diffusion graphs in the first field of the training set; the news diffusion graph is used to represent the diffusion path of news;

Specifically, the prior art uses human subjective consciousness to judge and detect the authenticity of news, which is inefficient. For example, as shown in FIG2 , in the embodiment of the present application, the domain with a relatively large amount of data is the source domain, and the domain with a relatively small amount of data is the target domain. The model trained by the source domain and the target domain can accurately identify fake news in the first domain (with only one error), while the detection results of fake news in the second domain will have a large number of errors (with 6 errors).

In order to solve the above problems, in an embodiment of the present application, multiple news diffusion graphs of the first field in the training set are first obtained; wherein the news diffusion graph is used to represent the diffusion path of news; optionally, the diffusion of news in social media mainly depends on the behavior of users in social media, and users' reposting, commenting or liking of news in social media constitutes a social media news diffusion graph; that is, users' reposting, commenting or liking of news are used as nodes in the news diffusion graph, and according to the nodes and time information in the news diffusion graph, a news diffusion graph can be constructed, thereby accurately representing the diffusion path of news.

Step 102: obtaining a plurality of news diffusion graphs in the second field in the training set; the ratio of the number of news diffusion graphs in the second field to the number of news diffusion graphs in the first field is less than a threshold;

Specifically, when an emergency occurs, that is, when a new domain appears, how to accurately detect the authenticity of the news in the newly emerged domain in the case of little data, so as to take timely measures to avoid the further spread of fake news, is a problem that technicians in this field need to solve urgently. In the embodiment of the present application, after obtaining multiple news propagation graphs in the first field, multiple news propagation graphs in the second field in the training set are also obtained, wherein the number of news propagation graphs in the second field is much smaller than the number of news propagation graphs in the first field; for example, the ratio of the number of news propagation graphs in the second field to the number of news propagation graphs in the first field is less than the threshold; for example, the number of news propagation graphs in the second field is 10% of the number of news propagation graphs in the first field; that is, in the process of training the social media fake news detection model, the present application not only obtains news in the first field with a large amount of data and a long propagation time as training samples, but also obtains news in the second field with a small amount of data and a short propagation time as training samples, so that the trained social media fake news detection model can not only accurately detect the authenticity of news with a large amount of data and a long propagation time, but also accurately detect the authenticity of social media news with a small amount of data and a short propagation time that has just occurred.

Step 103: According to multiple news propagation graphs in the first field, multiple news propagation graphs in the second field and the first target loss function, a social media fake news detection model is trained to obtain a social media fake news detection model trained based on a training set; the social media fake news detection model is used to detect the authenticity of news; the first target loss function is determined by classification loss, global contrast loss and local contrast loss; the global contrast loss represents the degree of correlation between node features in the news propagation graph, node features in the first type of augmented graph of the news propagation graph and news propagation graph features; the local contrast loss represents the degree of correlation between node features in the second type of augmented graph of the news propagation graph and node features in the third type of augmented graph of the news propagation graph; the classification loss represents the accuracy of the classification result.

Specifically, in the embodiment of the present application, after obtaining multiple news diffusion graphs in the first field with a large amount of data and a long diffusion time, and the news diffusion graphs in the second field with a small amount of data and a short diffusion time, the social media fake news detection model is trained according to the multiple news diffusion graphs in the first field, the multiple news diffusion graphs in the second field and the first target loss function, to obtain a social media fake news detection model trained based on the training set; wherein the social media fake news detection model is used to detect the authenticity of the news. Optionally, the first target loss function is determined by classification loss, global contrast loss and local contrast loss; wherein the global contrast loss is used to represent the degree of association between node features in the news diffusion graph, node features in the first type of augmented graph of the news diffusion graph and news diffusion graph features; the local contrast loss is used to represent the degree of association between node features in the second type of augmented graph of the news diffusion graph and node features in the third type of augmented graph of the news diffusion graph; that is, the global contrast loss is used to represent the degree of association between the node and the news diffusion graph, and the local contrast loss is used to represent the degree of association between nodes and nodes. Optionally, the global contrast loss can be compared and learned based on the original news propagation graph and the augmented graph obtained by randomly shuffling the nodes of the original news propagation graph, to obtain the relationship characteristics between each node in the original news propagation graph and the entire graph; the local contrast loss can be compared and learned based on the two types of augmented graphs obtained by selectively discarding the edges of the original news propagation graph and selectively covering the node features of the original graph to obtain the relationship characteristics between each node in the graph. That is, in the embodiment of the present application, by adding global contrast loss and local contrast loss, the optimization of the feature extractor in the social media fake news detection model is achieved, so that the optimized feature extractor and the social media fake news detection model can more accurately and comprehensively extract the features in the news propagation graph, and extract the features of fake news more accurately and completely, so that more and more important information can be extracted for news classification, and thus the authenticity of the news can be better and more accurately classified. On the other hand, the training method of the social media fake news detection model in the embodiment of the present invention performs local contrast learning and global contrast learning on the fake news data and its augmented data in the second field, learns the local information and global information of the fake news data in the second field, and improves the generalization ability of the model on the fake news data in the second field.

That is, after the embodiment of the present application trains the social media fake news detection model based on the first objective loss function, the trained social media fake news detection model can not only accurately extract the feature information of news in the first field and the global contrast loss between the nodes in the news propagation graph and the news propagation graph, and the local contrast loss between the augmented graph nodes in the news propagation graph, but also accurately extract the feature information of news in the second field and the global contrast loss between the nodes in the news propagation graph and the news propagation graph, and the local contrast loss between the augmented graph nodes in the news propagation graph. This means that the social media fake news detection model can accurately identify and extract the feature information of fake news, effectively promote and improve the detection accuracy of fake news, and thus the trained social media fake news detection model can accurately detect the authenticity of news in the second field with less data volume and shorter propagation time, thereby improving the accuracy of fake news detection in the second field with less data volume and shorter propagation time.

The method of the above embodiment trains a social media fake news detection model according to multiple news propagation graphs in the first field, multiple news propagation graphs in the second field and a first target loss function, so that the trained social media fake news detection model can not only accurately extract the feature information of news in the first field and the global contrast loss between the nodes in the news propagation graph and the news propagation graph, and the local contrast loss between the augmented graph nodes in the news propagation graph, but also accurately extract the feature information of news in the second field and the global contrast loss between the nodes in the news propagation graph and the news propagation graph, and the local contrast loss between the augmented graph nodes in the news propagation graph. This means that the trained social media fake news detection model can accurately and comprehensively extract the feature information of fake news in the second field with less data volume and shorter propagation time, effectively promoting and improving the detection accuracy of fake news, so that the trained social media fake news detection model can accurately detect the authenticity of news in the second field with less data volume and shorter propagation time, thereby improving the accuracy of fake news detection in the second field with less data volume and shorter propagation time.

In one embodiment, a social media fake news detection model includes at least one of the following:

Feature extraction module: The feature extraction module is used to extract the features of the news dissemination graph;

Classification module: The classification module is used to predict the authenticity of social media news corresponding to the news diffusion graph based on the characteristics of the news diffusion graph;

Self-supervised learning module; the self-supervised learning module is used to determine the global contrast loss based on the node features in the news diffusion graph, the node features in the first type of augmented graph of the news diffusion graph, and the news diffusion graph features; and determine the local contrast loss based on the node features in the second type of augmented graph of the news diffusion graph and the node features in the third type of augmented graph of the news diffusion graph.

Specifically, in an embodiment of the present application, a social media fake news detection model includes a feature extraction module, a classification module and a self-supervised learning module; wherein the feature extraction module is used to extract news propagation graph features; optionally, the feature extraction module is established based on a graph convolutional network (GCN); the classification module is used to predict the authenticity of social media news corresponding to the news propagation graph based on the news propagation graph features; the self-supervised learning module is used to determine the global contrast loss based on the node features in the news propagation graph, the node features in the first type of augmented graph of the news propagation graph, and the news propagation graph features; the local contrast loss is determined based on the node features in the second type of augmented graph of the news propagation graph and the node features in the third type of augmented graph of the news propagation graph, thereby enabling the social media fake news detection model to not only accurately extract the feature information of the news in the first field and the global contrast between the nodes in the news propagation graph and the news propagation graph The contrast loss, the local contrast loss between the augmented graph nodes in the news propagation graph, and the feature information of the news in the second field can also be accurately extracted, as well as the global contrast loss between the nodes in the news propagation graph and the news propagation graph, and the local contrast loss between the augmented graph nodes in the news propagation graph; optionally, the global contrast loss and local contrast loss corresponding to fake news are different from the global contrast loss and local contrast loss corresponding to true news; the global contrast loss and local contrast loss corresponding to fake news in the first field and the global contrast loss and local contrast loss corresponding to fake news in the second field are also different, so that the trained social media fake news detection model can accurately identify and extract the feature information of fake news, realize the detection of the authenticity of the second field news with less data volume and shorter propagation time, and improve the accuracy of detecting fake news in the second field with less data volume and shorter propagation time.

In one embodiment, the social media fake news detection model is trained based on the following method:

Input multiple news propagation graphs in the first domain and multiple news propagation graphs in the second domain in the training set into the social media fake news detection model, and output the authenticity detection results of the social media news corresponding to the news propagation graphs; obtain the classification loss of the social media fake news detection model based on the authenticity detection results of the social media news and the label information of the news; the label information is used to mark the authenticity of the news;

Input multiple news propagation graphs in the first field in the training set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the first field, the second type of augmented graphs corresponding to the multiple news propagation graphs in the first field, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the first field into the social media fake news detection model to obtain the global contrast loss and local contrast loss of the multiple news in the first field;

Input multiple news propagation graphs in the second field in the training set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the second field, the second type of augmented graphs corresponding to the multiple news propagation graphs in the second field, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the second field into the social media fake news detection model to obtain the global contrast loss and local contrast loss of the multiple news in the second field;

The weighted sum of the classification loss of the social media fake news detection model, the global contrast loss and local contrast loss of multiple news in the first field, and the global contrast loss and local contrast loss of multiple news in the second field is taken as the value of the first objective loss function;

Based on the value of the first objective loss function, the social media fake news detection model is trained to obtain a social media fake news detection model trained based on the training set.

Specifically, in the process of training the social media fake news detection model, the present application first inputs multiple news propagation graphs in the first field and multiple news propagation graphs in the second field in the training set into the social media fake news detection model, and the classification module of the social media fake news detection model outputs the authenticity detection results of the social media news corresponding to the news propagation graphs. Then, based on the authenticity detection results of the social media news and the label information of the news, the classification loss (detection accuracy) of the social media fake news detection model can be obtained.

Furthermore, in an embodiment of the present application, multiple news propagation graphs in the first field in the training set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the first field, the second type of augmented graphs corresponding to the multiple news propagation graphs in the first field, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the first field are input into the social media fake news detection model to obtain the global contrast loss and local contrast loss of the multiple news in the first field; that is, the self-supervised learning module in the social media fake news detection model obtains the local contrast loss and the global contrast loss of the nodes in the graph according to the features of the original news propagation graph and its three types of augmented graphs; wherein the global contrast loss is based on the original news propagation graph and the augmented graph obtained by randomly shuffling the nodes of the original news propagation graph, and comparative learning is performed to obtain the relationship features between each node in the original news propagation graph and the entire graph; the local contrast loss is based on the relationship features between each node in the graph obtained by comparative learning of two types of augmented graphs obtained by selectively discarding the edges of the original news propagation graph and selectively covering the node features of the original graph.

For example, the global contrast loss between each node in the original news propagation graph and the entire graph is modeled using the following formula:

D(Z _si ,s)＝Sigmoid(Z _si *s)

Among them, Z _si represents the node feature representation of the news diffusion graph; s represents the feature representation of the news diffusion graph; * represents the inner product; D represents the discriminator, which calculates the correlation scores of positive and negative samples respectively; Sigmoid is the activation function.

The global contrast loss between each node in the augmented graph obtained by randomly shuffling the nodes of the original news diffusion graph and the entire graph is determined based on the following formula:

Where N represents the number of nodes in the input graph; Z _0i represents the nodes in the original news diffusion graph; and Z _1i represents the nodes in the augmented graph obtained by randomly shuffling the nodes of the original news diffusion graph.

Optionally, the local contrast loss is modeled using the following formula:

Wherein, Z ₂ and Z ₃ represent two types of augmented graphs of news data; optionally, they may be two types of augmented graphs obtained by selectively discarding the edges of the original news diffusion graph and selectively covering the node features of the original graph;

(Z _2i ,Z _3j )(i,j∈{1,...,N},i≠j), N represents the number of nodes in the graph, where cos() represents the cosine similarity function, τ is a hyperparameter, and g() is used to further enhance the representation ability of the model. The nodes calculated by g() are represented as Z2' and Z3';

Where I represents the identity matrix, and the local contrast loss of the node is finally defined as follows:

Furthermore, multiple news propagation graphs in the second field in the training set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the second field, the second type of augmented graphs corresponding to the multiple news propagation graphs in the second field, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the second field are input into the social media fake news detection model to obtain the global contrast loss and local contrast loss of the multiple news in the second field; optionally, the global contrast loss and local contrast loss of the multiple news in the second field are consistent with the method for determining the global contrast loss and the local contrast loss in the first field, and the embodiments of the present application will not be repeated.

Optionally, after determining the classification loss of the social media fake news detection model, the global contrast loss and local contrast loss of multiple news in the first field, and the global contrast loss and local contrast loss of multiple news in the second field, in an embodiment of the present application, the weighted sum of the classification loss of the social media fake news detection model, the global contrast loss and local contrast loss of multiple news in the first field, and the global contrast loss and local contrast loss of multiple news in the second field is used as the value of the first target loss function; that is, in the process of training the social media fake news detection model, the embodiment of the present application not only simply considers the classification loss of the social media fake news detection model, but also considers whether the social media fake news detection model can accurately and comprehensively extract the feature information of the fake news in the second field with a small amount of data and a short propagation time, so that after accurately extracting the feature information of the fake news in the second field with a small amount of data and a short propagation time, it can also effectively promote the recognition accuracy of fake news, fundamentally solve the problem of not being able to accurately and comprehensively extract the feature information of the fake news in the second field with a small amount of data and a short propagation time, and effectively improve the accuracy of fake news detection in the second field with a small amount of data and a short propagation time.

The method of the above embodiment, during the process of training the social media fake news detection model, not only simply considers the classification loss of the social media fake news detection model, but also considers whether the social media fake news detection model can accurately and comprehensively extract the feature information of fake news in the second field with less data volume and shorter propagation time. Therefore, after accurately extracting the feature information of fake news in the second field with less data volume and shorter propagation time, the accuracy of fake news recognition can be effectively promoted, fundamentally solving the problem of not being able to accurately and comprehensively extract the feature information of fake news in the second field with less data volume and shorter propagation time, and effectively improving the accuracy of fake news detection in the second field with less data volume and shorter propagation time.

In one embodiment, the social media fake news detection model further includes:

Data adaptive constraint module: The data adaptive constraint module is used to determine the difference between the news diffusion graph features in the training set and the news diffusion graph features in the test set.

Specifically, in the embodiment of the present application, the social media fake news detection model, on the basis of including a feature extraction module, a classification module and a self-supervised learning module, further adds a data adaptive constraint module. Among them, the data adaptive constraint module is used to determine the difference between the news propagation graph features in the training set and the news propagation graph features in the test set. That is, after completing the training of the social media fake news detection model based on the training set, the embodiment of the present application determines the difference between the news propagation graph features in the training set and the news propagation graph features in the test set based on the data adaptive constraint module, thereby achieving fine-tuning of the social media fake news detection model, so that the performance of the feature extractor on the test set is similar to that of the feature extractor on the training set, and the feature extractor is optimized to prevent the occurrence of overfitting problems, so that the social media fake news detection model after constraining the difference between the news propagation graph features in the training set and the news propagation graph features in the test set by adding a data adaptive constraint module can more accurately identify the fake news in the second field to be detected.

For example, the difference between the news diffusion graph features of the second field in the training set and the news diffusion graph features of the second field in the test set can be determined based on the following method:

Among them, h is the feature matrix of the training set data, μ represents the feature mean, Σ represents the covariance matrix, and the feature mean μ _t and covariance matrix Σ _t of the test set are obtained in the same way. The difference between the news diffusion graph features of the second field in the training set and the news diffusion graph features of the second field in the test set is as follows:

That is, the data adaptive constraint module in the embodiment of the present application ensures that the social media fake news detection model does not overfit the fake news data by calculating the difference between the two statistical values of the fake news data in the training set and the fake news data in the test set, thereby improving the generalization ability of the model.

In the method of the above embodiment, the social media fake news detection model constrains the difference between the news propagation graph features in the training set and the news propagation graph features in the test set by adding a data adaptive constraint module, thereby achieving fine-tuning of the social media fake news detection model, so that the performance of the feature extractor on the test set is similar to that of the feature extractor on the training set, thereby achieving optimization of the feature extractor and preventing the occurrence of overfitting problems. Therefore, the social media fake news detection model after constraining the difference between the news propagation graph features in the training set and the news propagation graph features in the test set by adding a data adaptive constraint module can more accurately identify fake news in the second field to be detected.

In one embodiment, after training the social media fake news detection model based on the value of the first objective loss function to obtain the social media fake news detection model trained based on the training set, the method further includes:

Input multiple news propagation graphs in the second field in the test set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set, the second type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set into the social media fake news detection model trained based on the training set, and obtain the global contrast loss and local contrast loss of the multiple news in the second field in the test set;

According to the difference between the news propagation graph features of the second field in the training set and the news propagation graph features of the second field in the test set, the weighted sum of the global contrast loss and the local contrast loss of multiple news in the second field in the test set is used as the value of the second objective loss function;

Based on the value of the second objective loss function, the social media fake news detection model is trained to obtain a social media fake news detection model based on the potential features of the test set.

Specifically, the embodiment of the present application inputs multiple news propagation graphs in the second field in the test set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set, the second type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set into a social media fake news detection model trained based on the training set to obtain the global contrast loss and local contrast loss of the multiple news in the second field in the test set; optionally, the method steps for determining the global contrast loss and local contrast loss of the multiple news in the second field in the test set are similar to the method steps for determining the global contrast loss and local contrast loss of the multiple news in the training set, and the embodiments of the present application will not be repeated herein.

Furthermore, after obtaining the global contrast loss and local contrast loss of multiple news in the second field in the test set, the difference between the news propagation graph features of the second field in the training set and the news propagation graph features of the second field in the test set can be determined based on the data adaptive constraint module in the social media fake news detection model; and then the difference between the news propagation graph features of the second field in the training set and the news propagation graph features of the second field in the test set, and the weighted sum of the global contrast loss and the local contrast loss of multiple news in the second field in the test set can be used as the value of the second objective loss function, and based on the value of the second objective loss function, the social media fake news detection model is trained to obtain a social media fake news detection model based on the potential features of the test set. That is, the constraint of the difference between the news propagation graph features in the second field in the training set and the news propagation graph features in the second field in the test set is added to the loss function of the training set, thereby achieving fine-tuning of the social media fake news detection model, making the performance of the feature extractor on the test set and the feature extractor on the training set similar, achieving optimization of the feature extractor, and preventing the occurrence of overfitting problems. Therefore, by adding the constraint on the difference between the news propagation graph features in the training set and the news propagation graph features in the test set, the social media fake news detection model can more accurately identify the fake news in the second field to be detected. Optionally, if the constraint of the difference between the news propagation graph features in the second field in the training set and the news propagation graph features in the second field in the test set is not added to the loss function, it will lead to the occurrence of overfitting problems, so that the social media fake news detection model trained based on the training set cannot accurately detect the authenticity of the news in the second field to be detected, and cannot accurately and effectively identify the fake news in the second field.

The method of the above embodiment adds a constraint on the difference between the news propagation graph features of the second field in the training set and the news propagation graph features of the second field in the test set to the loss function, thereby achieving fine-tuning of the social media fake news detection model and optimization of the feature extractor, so that the performance of the feature extractor on the test set is similar to that of the feature extractor on the training set, effectively preventing the occurrence of overfitting problems, and thus making the social media fake news detection model after adding the constraint of the difference between the news propagation graph features in the training set and the news propagation graph features in the test set can more accurately identify fake news in the second field to be detected.

The present application embodiment provides a method for detecting fake news on social media, including:

Obtain a news dissemination graph of a second field to be detected;

The news propagation graph of the second field to be detected is input into the social media fake news detection model to obtain the authenticity detection result of the social media news corresponding to the news propagation graph of the second field; the social media fake news detection model is trained based on the training method of the social media fake news detection model such as any of the above items.

Specifically, after training the social media fake news detection model based on the training set and the test set, the fake news in the second field can be accurately detected based on the trained social media fake news detection model. Optionally, the news propagation graph of the second field to be detected can be first obtained, and then the news propagation graph of the second field to be detected can be input into the trained social media fake news detection model, and the authenticity detection result of the social media news corresponding to the news propagation graph of the second field can be obtained, thereby realizing the authenticity detection of the second field news with less data volume and shorter propagation time, and improving the accuracy of the second field fake news detection with less data volume and shorter propagation time.

Exemplarily, the specific process of the training method of the social media fake news detection model in the embodiment of the present application is shown in Figures 3 and 4, and is as follows:

1. Train the social media fake news detection model based on the first domain (high resource) news in the training set.

By performing three different types of augmentation operations on the first domain news data: edge loss, feature shuffling, and feature masking, three different types of augmented graphs are generated. A feature extractor is used to obtain the high-dimensional feature matrix of the first domain news data and the three types of augmented graphs. The main task and the auxiliary task share a feature extractor. The first domain fake news data is input into the main task, and the first domain news data and the three types of augmented graphs are input into the auxiliary task. The augmented graphs with edge loss and feature random masking are locally compared and learned, and the first domain news data and the augmented graphs with feature shuffling are globally compared and learned.

1.1 Global Contrastive Learning Based on Graph Convolutional Neural Networks

The purpose of global contrastive learning is to help the node representation in the graph data obtain the global information of the entire graph. Given an input event Different graphs can be generated through various types of data augmentation. For global contrastive learning, two views are used: one is the original view View0; the other is the enhanced view View1, where the node attributes of all nodes in the graph are randomly assigned. With these two graphs, two corresponding node representations Z0 and Z1 can be obtained through a shared feature extractor. Afterwards, a global graph representation s can be summarized from the node representation matrix Z0 of the original view View0 through a multi-layer perceptron.

The positive samples in global contrastive learning are composed of node-graph representation pairs, where both the node representation and the graph representation come from the original view View0. Negative samples are also composed of node-graph representation pairs, where the node representation comes from View1 and the graph representation comes from View0. The discriminator D calculates the scores of positive and negative samples respectively. For positive samples, the score should be higher, and for negative samples, the score should be lower. The representation of D is defined as follows:

D(Z _si ,s)＝Sigmoid(Z _si *s)

Where Z _si represents the node representation of the graph Views, and * represents the inner product. The loss function of global contrastive learning is as follows:

Where N represents the number of nodes in the input graph.

1.2 Local Contrastive Learning Based on Graph Convolutional Neural Networks

Through local contrastive learning, the model can learn richer node feature representations. By comparing the feature differences between a node and its neighboring nodes, the model can learn more discriminative node representations, which can reduce the impact of node position, noise, and missing values on graph data, and improve the robustness and interpretability of graph data. Through local contrastive learning, the model can better capture the similarities and differences between a node and its neighboring nodes, thereby better understanding the contextual information of the node.

Given an input event Through the graph augmentation strategy (edge deletion and node attribute masking), two different augmented graphs are obtained as the input of the feature extractor. The output of the feature extractor is two node representation matrices, Z ₂ and Z ₃ . The basic goal of local contrastive learning is to distinguish whether two nodes from the enhanced view are the same node. Therefore, (Z _2i ,Z _3i )(i∈{1,...,N}) represents a positive pair, where N is the number of nodes, (Z _2i ,Z _3j ) and (Z _2i ,Z _2j )(i,j∈{1,...,N},i≠j) represent intra-class negative pairs and inter-class negative pairs. The objective function of the positive pair is defined as follows:

Where cos() represents the cosine similarity function, τ is a hyperparameter, and g() is a two-layer MLP. The nodes that pass through the MLP are represented as Z2' and Z3', and the regularizer is used on the redefined node representation. As follows:

The final loss function of local comparison is defined as follows:

Based on the training method of high-resource fake news data trained at test time, the final loss function of the auxiliary task is as follows:

_Ls ＝ _Lg + _αLl

Based on the training method of high-resource fake news data trained at test time, the loss function is as follows:

L＝ _Lm + _γLs

Where Lm is the loss of the main task.

2. Train the social media fake news detection model based on the second domain (low-resource) news in the training set.

Perform three different types of augmentation operations on the second domain news data: edge loss, feature shuffling, and feature masking to generate three different types of augmented graphs. Use a feature extractor to obtain the high-dimensional feature matrix of the second domain news data and the three types of augmented graphs. The feature extractor here is the same as the feature extractor mentioned in the first part. Input the second domain news data and the three types of augmented graphs into the auxiliary task, perform local contrast learning on the augmented graphs that have undergone edge loss and feature random masking operations, and perform global contrast learning on the second domain fake news data and the augmented graphs that have undergone feature shuffling operations. The training of the social media fake news detection model based on the second domain news in the training set also adopts a self-supervised learning framework that includes global contrast learning and local contrast learning. The model of the auxiliary task is the same as the training method of the social media fake news detection model based on the first domain news in the training set, so it will not be repeated here.

3. Data adaptive constraint method based on training during test time.

Use the READOUT function to obtain the feature matrix of the news data, and calculate two statistics of the news data feature matrix, the feature mean and the covariance matrix. Calculate two statistics of the fake news data in the second field one by one, the eigenvalue and the covariance matrix, and construct the loss function.

definition Use the READOUT function (READOUT function is a feature extractor)

Get the feature matrix of news data, calculate two statistical data of the feature matrix of news data, the feature mean and covariance matrix, the definitions of the feature mean and covariance matrix are as follows:

The feature matrix of the low-resource fake news data in the second domain is calculated through the READOUT function. Its two statistical data, the feature mean and the covariance matrix, are also calculated, and the loss function is constructed as follows:

where μ _t ,∑ _t represent the feature mean and covariance matrix of low-resource fake news data, respectively.

The method of the above embodiment determines whether the news data is fake news by obtaining the news data to be detected and inputting the news data into the fake news detection model. The model training method of the present invention integrates the fake news data features of two different fields into the model representation, thereby achieving better modeling of the target data domain, and training the target data again in the test phase, thereby enhancing the model's ability to represent the target data, and can detect fake news more accurately and efficiently, thereby improving the accuracy of cross-domain detection of fake news.

The training device for the social media fake news detection model provided by the present invention is described below. The training device for the social media fake news detection model described below and the training method for the social media fake news detection model described above can refer to each other.

FIG5 is a schematic diagram of the structure of a training device for a social media fake news detection model provided by the present invention. The training device for a social media fake news detection model provided by this embodiment includes:

The first acquisition module 710 is used to acquire a plurality of news diffusion graphs in the first field in the training set; the news diffusion graph is used to represent the diffusion path of news;

The second acquisition module 720 is used to acquire a plurality of news diffusion graphs in a second field in the training set; the ratio of the number of news diffusion graphs in the second field to the number of news diffusion graphs in the first field is less than a threshold;

The training module 730 is used to train the social media fake news detection model according to multiple news propagation graphs in the first field, multiple news propagation graphs in the second field and the first target loss function, so as to obtain the social media fake news detection model trained based on the training set; the social media fake news detection model is used to detect the authenticity of news; the first target loss function is determined by classification loss, global contrast loss and local contrast loss; the global contrast loss represents the degree of correlation between the node features in the news propagation graph, the node features in the first type of augmented graph of the news propagation graph and the features of the news propagation graph; the local contrast loss represents the degree of correlation between the node features in the second type of augmented graph of the news propagation graph and the node features in the third type of augmented graph of the news propagation graph; the classification loss represents the accuracy of the classification result.

Optionally, the social media fake news detection model includes at least one of the following:

Feature extraction module: The feature extraction module is used to extract the features of the news dissemination graph;

Classification module: The classification module is used to predict the authenticity of social media news corresponding to the news diffusion graph based on the characteristics of the news diffusion graph;

Self-supervised learning module; the self-supervised learning module is used to determine the global contrast loss based on the node features in the news diffusion graph, the node features in the first type of augmented graph of the news diffusion graph, and the news diffusion graph features; and determine the local contrast loss based on the node features in the second type of augmented graph of the news diffusion graph and the node features in the third type of augmented graph of the news diffusion graph.

Optionally, the social media fake news detection model is trained based on:

Input multiple news propagation graphs in the first domain and multiple news propagation graphs in the second domain in the training set into the social media fake news detection model, and output the authenticity detection results of the social media news corresponding to the news propagation graphs; obtain the classification loss of the social media fake news detection model based on the authenticity detection results of the social media news and the label information of the news; the label information is used to mark the authenticity of the news;

Input multiple news propagation graphs in the first field in the training set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the first field, the second type of augmented graphs corresponding to the multiple news propagation graphs in the first field, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the first field into the social media fake news detection model to obtain the global contrast loss and local contrast loss of the multiple news in the first field;

Input multiple news propagation graphs in the second field in the training set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the second field, the second type of augmented graphs corresponding to the multiple news propagation graphs in the second field, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the second field into the social media fake news detection model to obtain the global contrast loss and local contrast loss of the multiple news in the second field;

The weighted sum of the classification loss of the social media fake news detection model, the global contrast loss and local contrast loss of multiple news in the first field, and the global contrast loss and local contrast loss of multiple news in the second field is taken as the value of the first objective loss function;

Based on the value of the first objective loss function, the social media fake news detection model is trained to obtain a social media fake news detection model trained based on the training set.

Optionally, the social media fake news detection model also includes:

Data adaptive constraint module: The data adaptive constraint module is used to determine the difference between the news diffusion graph features in the training set and the news diffusion graph features in the test set.

Optionally, the training module 730 is further used to input multiple news propagation graphs in the second field in the test set, the first type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set, the second type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set, and the third type of augmented graphs corresponding to the multiple news propagation graphs in the second field in the test set into the social media fake news detection model trained based on the training set, so as to obtain the global contrast loss and local contrast loss of the multiple news in the second field in the test set;

According to the difference between the news propagation graph features of the second field in the training set and the news propagation graph features of the second field in the test set, the weighted sum of the global contrast loss and the local contrast loss of multiple news in the second field in the test set is used as the value of the second objective loss function;

Based on the value of the second objective loss function, the social media fake news detection model is trained to obtain a social media fake news detection model based on the potential features of the test set.

The device of the embodiment of the present invention is used to execute the method in any of the aforementioned method embodiments. Its implementation principle and technical effects are similar and will not be repeated here.

FIG6 illustrates a schematic diagram of the physical structure of an electronic device, which may include: a processor 810, a communications interface 820, a memory 830 and a communications bus 840, wherein the processor 810, the communications interface 820, and the memory 830 communicate with each other via the communications bus 840. The processor 810 can call the logic instructions in the memory 830 to execute the training method of the social media fake news detection model, which includes: obtaining multiple news propagation graphs in the first field in the training set; the news propagation graph is used to represent the propagation path of news; obtaining multiple news propagation graphs in the second field in the training set; the ratio of the number of news propagation graphs in the second field to the number of news propagation graphs in the first field is less than a threshold; according to the multiple news propagation graphs in the first field, the multiple news propagation graphs in the second field and the first target loss function, the social media fake news detection model is trained to obtain the social media fake news detection model based on the training set; the social media fake news detection model is used to detect the authenticity of news; the first target loss function is determined by classification loss, global contrast loss and local contrast loss; the global contrast loss represents the degree of correlation between the node features in the news propagation graph, the node features in the first type of augmented graph of the news propagation graph and the news propagation graph features; the local contrast loss represents the degree of correlation between the node features in the second type of augmented graph of the news propagation graph and the node features in the third type of augmented graph of the news propagation graph; the classification loss represents the accuracy of the classification result.

In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the training method of the social media fake news detection model provided by the above-mentioned methods, the method comprising: obtaining multiple news propagation graphs in a first field in a training set; the news propagation graphs are used to represent the propagation path of news; obtaining multiple news propagation graphs in a second field in the training set; the ratio of the number of news propagation graphs in the second field to the number of news propagation graphs in the first field is less than a threshold value; based on the multiple news propagation graphs in the first field and the multiple news propagation graphs in the second field, The news propagation graph and the first objective loss function are used to train the social media fake news detection model to obtain the social media fake news detection model trained based on the training set; the social media fake news detection model is used to detect the authenticity of news; the first objective loss function is determined by classification loss, global contrast loss and local contrast loss; the global contrast loss represents the degree of association between the node features in the news propagation graph, the node features in the first type of augmented graph of the news propagation graph and the features of the news propagation graph; the local contrast loss represents the degree of association between the node features in the second type of augmented graph of the news propagation graph and the node features in the third type of augmented graph of the news propagation graph; the classification loss represents the correctness of the classification result.

On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the training methods for the social media fake news detection models provided above.

The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or 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. Those of ordinary skill in the art may understand and implement it without creative effort.

Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM/RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (10)

1. The training method of the social media false news detection model is characterized by comprising the following steps of:

acquiring a plurality of news propagation graphs of a first field in a training set; the news propagation graph is used for representing a news propagation path;

acquiring a plurality of news propagation graphs of a second field in the training set; the ratio of the number of the news broadcasting pictures in the second field to the number of the news broadcasting pictures in the first field is smaller than a threshold value;

training the social media false news detection model according to the plurality of news propagation graphs in the first field, the plurality of news propagation graphs in the second field and the first target loss function to obtain a social media false news detection model trained based on a training set; the social media false news detection model is used for detecting the authenticity of news; the first objective loss function is determined from classification loss, global contrast loss, and local contrast loss; the global contrast loss represents the node characteristics in the news propagation graph, and the degree of association between the node characteristics in the first type of the news propagation graph and the news propagation graph characteristics; the local contrast loss represents a degree of association between node features in a second type of augmentation graph of the news propagation graph and node features in a third type of augmentation graph of the news propagation graph; the classification loss represents the accuracy of the classification result.

2. The method of training a social media false news detection model according to claim 1, wherein the social media false news detection model comprises at least one of:

a feature extraction module; the feature extraction module is used for extracting news propagation graph features;

a classification module; the classification module is used for predicting the authenticity of social media news corresponding to the news propagation graph according to the news propagation graph characteristics;

a self-supervision learning module; the self-supervision learning module is used for determining the global contrast loss according to the node characteristics in the news propagation graph, the node characteristics in the first type of the news propagation graph and the news propagation graph characteristics; and determining the local contrast loss according to the node characteristics in the second type augmentation chart of the news propagation chart and the node characteristics in the third type augmentation chart of the news propagation chart.

3. The training method of a social media false news detection model according to claim 2, wherein the social media false news detection model is trained based on the following manner:

inputting a plurality of news propagation graphs of a first field and a plurality of news propagation graphs of a second field in a training set into a social media false news detection model, and outputting an authenticity detection result of social media news corresponding to the news propagation graphs; obtaining the classification loss of the social media false news detection model according to the authenticity detection result of the social media news and the label information of the news; the tag information is used for marking the authenticity of the news;

Inputting a plurality of news propagation graphs of the first field, a first type of augmentation graph corresponding to the plurality of news propagation graphs of the first field, a second type of augmentation graph corresponding to the plurality of news propagation graphs of the first field and a third type of augmentation graph corresponding to the plurality of news propagation graphs of the first field in a training set to a social media false news detection model to obtain global contrast loss and local contrast loss of a plurality of news of the first field;

inputting a plurality of news propagation graphs of the second field, a first type of augmentation graph corresponding to the plurality of news propagation graphs of the second field, a second type of augmentation graph corresponding to the plurality of news propagation graphs of the second field and a third type of augmentation graph corresponding to the plurality of news propagation graphs of the second field in a training set to a social media false news detection model to obtain global contrast loss and local contrast loss of a plurality of news of the second field;

taking the weighted sum of the classification loss of the social media false news detection model, the global contrast loss and the local contrast loss of a plurality of news in the first field, and the global contrast loss and the local contrast loss of a plurality of news in the second field as the value of a first target loss function;

And training the social media false news detection model based on the value of the first target loss function to obtain a social media false news detection model trained based on a training set.

4. The method for training a social media false news detection model according to claim 3, further comprising:

a data self-adaptive constraint module; the data adaptive constraint module is used for determining differences between the news spread map features in the training set and the news spread map features in the test set.

5. The method for training a social media false news detection model according to claim 4, wherein training the social media false news detection model based on the value of the first objective loss function, after obtaining a social media false news detection model trained based on a training set, further comprises:

inputting a plurality of news propagation graphs in a second field in a test set, a first type augmentation graph corresponding to the plurality of news propagation graphs in the second field in the test set, a second type augmentation graph corresponding to the plurality of news propagation graphs in the second field in the test set and a third type augmentation graph corresponding to the plurality of news propagation graphs in the second field in the test set into the social media false news detection model trained based on the training set to obtain global contrast loss and local contrast loss of a plurality of news in the second field in the test set;

Taking the sum of the global contrast loss and the local contrast loss weights of a plurality of news in the second field in the test set as a value of a second target loss function according to the difference between the news propagation map features in the second field in the training set and the news propagation map features in the second field in the test set;

and training the social media false news detection model based on the value of the second target loss function to obtain the social media false news detection model based on the potential characteristics of the test set.

6. A method for detecting false news of social media, comprising:

acquiring a news propagation diagram of a second field to be detected;

inputting the news propagation diagram of the second field to be detected into the social media false news detection model to obtain an authenticity detection result of social media news corresponding to the news propagation diagram of the second field; the social media false news detection model is trained based on the method of any one of claims 1-5.

7. A training device for a social media false news detection model, comprising:

the first acquisition module is used for acquiring a plurality of news propagation graphs of a first field in the training set; the news propagation graph is used for representing a news propagation path;

The second acquisition module is used for acquiring a plurality of news propagation graphs in a second field in the training set; the ratio of the number of the news broadcasting pictures in the second field to the number of the news broadcasting pictures in the first field is smaller than a threshold value;

the training module is used for training the social media false news detection model according to the plurality of news propagation graphs in the first field, the plurality of news propagation graphs in the second field and the first target loss function to obtain a social media false news detection model trained based on a training set; the social media false news detection model is used for detecting the authenticity of news; the first objective loss function is determined from classification loss, global contrast loss, and local contrast loss; the global contrast loss represents the node characteristics in the news propagation graph, and the degree of association between the node characteristics in the first type of the news propagation graph and the news propagation graph characteristics; the local contrast loss represents a degree of association between node features in a second type of augmentation graph of the news propagation graph and node features in a third type of augmentation graph of the news propagation graph; the classification loss represents the accuracy of the classification result.

8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the training method of the social media false news detection model of any one of claims 1 to 5 or the social media false news detection method of claim 6 when the program is executed by the processor.

9. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor implements the training method of the social media false news detection model according to any one of claims 1 to 5 or the social media false news detection method according to claim 6.

10. A computer program product having stored thereon executable instructions that, when executed by a processor, cause the processor to implement the training method of a social media false news detection model according to any one of claims 1 to 5 or the social media false news detection method according to claim 6.