A blockchain phishing node detection method based on vertical federated learning

Abstract

The invention discloses a block chain phishing node detection method based on vertical federated learning, which is mainly applied to the detection of phishing nodes in a block chain transaction network. By querying the transaction records related to the target users, two small-scale sub-networks of time-series type and network-sequence type are constructed for training. Using the vertical federated learning strategy, the network embedding and original node class labels calculated by each edge-end phishing detection model are uploaded to the server, and the server-side combines the network embedding from the edge-end detection model through the attention mechanism. The server uses the combined network embedding to train its own detection model, and distributes the edge detection model to improve the performance of the blockchain phishing detection model at each edge.

Description

technical field [0001] The invention belongs to the field of block chain network security, and in particular relates to a block chain phishing node detection method based on vertical federated learning. Background technique [0002] Blockchain technology originated from Bitcoin. As an encrypted and secure electronic payment system, Bitcoin's original purpose was to solve the problem of over-reliance on trusted third parties in electronic payments. Over the past decade, cryptocurrencies in the financial sector, one of the most successful applications of blockchain technology, have become a breeding ground for new types of cybercrime. Although blockchain technology can guarantee the security of stored data and the reliability of transfer behavior, and store cryptocurrency in the user's own account, due to the lack of supervision by a third-party authority, once the transfer behavior to fraudsters is blocked Blockchain confirms that it is almost impossible for users to recover...

AI Technical Summary

Problems solved by technology

The difference is that in different blockchain phishing detection models, the same transaction data may be processed into network data in different formats, resulting in differences in network characteristics

Different blockchain phishing detection methods focus on different data characteristics (such as transaction time, transaction amount), and a single detection method may pay too much attention to a certain type of characteristics in transaction data, making it difficult to learn comprehensively Characteristics of a Potential Phisher

Images