Smart Learning to Find Dumb Contracts. (arXiv:2304.10726v1 [cs.CR])

We introduce Deep Learning Vulnerability Analyzer (DLVA), a vulnerability
detection tool for Ethereum smart contracts based on powerful deep learning
techniques for sequential data adapted for bytecode. We train DLVA to judge
bytecode even though the supervising oracle, Slither, can only judge source
code. DLVA's training algorithm is general: we "extend" a source code analysis
to bytecode without any manual feature engineering, predefined patterns, or
expert rules. DLVA's training algorithm is also robust: it overcame a 1.25%
error rate mislabeled …

algorithm analysis arxiv code code analysis data deep learning deep learning techniques detection engineering error ethereum expert feature feature engineering general oracle patterns rate rules smart smart contracts source code analysis tool training vulnerability vulnerable