For this project, I developed an algorithm for content-based file type detection that used clustering and byte-frequency signatures. This was implemented as an extension to Apache Tika . We also examined the effect of file size on normalized byte-frequency signatures. See the paper here
Inf is a first-order predicate logic inference engine that can build knowledge bases and query them using backward chaining.