In this paper, we explore the idea of using locality sensitive hashes as input features to a feed-forward neural network with the goal of detecting JavaScript malware through static analysis. An experiment is conducted using a dataset containing 1.5M …