Due to the variety of applications it can be used for and the potential for disruption, block chain technology has attracted a lot of attention. One of the main reasons block chain is so popular is due to its distinctive structure: it is described as a computerized log file and kept as a series of connected groups, or blocks, and employs peer-to-peer networks and registers to keep transactions. In this paper, we define an encoder-decoder deep learning model-based anomaly detection system that is trained using aggregate data acquired by observing block chain activity. The effectiveness of our methodology in identifying attacks that have been disclosed publicly has been demonstrated through experiments on the whole history logs of the Ethereum Classic network. As far as we are aware, our strategy is the first to offer a complete and workable method to monitor the security of block chain transactions.

Block chain, Attacks, Cyber security, Vulnerability, Anomaly detection