Gradient-based graph adversarial sample generation method by adding false nodes

Abstract

The invention discloses a gradient-based graph adversarial sample generation method by adding false nodes, which comprises the following steps: (1) acquiring original graph data and a graph node classification model, and constructing an attack target set; adding false nodes into the original image data to obtain adversarial samples; (2) selecting an attack target (v *, y *) from the attack targetset; (3) inputting a current adversarial sample into the classification model, and calculating loss by using a loss function; (4) calculating the gradient of the loss relative to the input adjacent matrix, and selecting the node corresponding to the element with the maximum gradient value in the row corresponding to the false node to be connected with the added false node to obtain a new adversarial sample; and (5) inputting the new adversarial sample into the classification model, if the classification result is y *, determining that the new adversarial sample is the generated adversarial sample, and otherwise, skipping to the step (3). By utilizing the method, the generated adversarial sample can effectively influence the classification result of the graph deep learning model, and the feasibility in practical application is higher.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence information security, and in particular relates to a gradient-based graph adversarial sample generation method by adding false nodes. Background technique [0002] A graph in graph theory is a graph composed of a number of given points and a line connecting two points. This graph is usually used to describe a certain relationship between certain things. Points represent things, and points connect two points. The line of represents that there is a certain relationship between the corresponding two things. The graph G in graph theory is an ordered pair (V, E), where V is called the vertex set, which is the set of all vertices in the graph, and E is called the edge set, which is the set of edges between all vertices. gather. Simply put, vertices represent things, and edges represent relationships between things. In addition, the attribute graph (Attributed Graph) can be expressed a...

AI Technical Summary

Problems solved by technology

In the field of graph data, current research is to make the classification results of target nodes mislead by adding or deleting existing edge or node features.

But this method may be difficult to implement in actual scenarios. For example, in a social network, if you want to delete or add an edge between two users, you may need to obtain the login permissions of these users, but this is difficult to obtain in actual situations.

Images