Toward the Discovery and Extraction of Money Laundering Evidence from Arbitrary Data Formats using Combinatory Reductions
Alonza Mumford, Duminda Wijesekera · 2014
Abstract—The evidence of money laundering schemes exist undetected in the electronic files of banks and insurance firms scattered around the world. Intelligence and law enforcement analysts, impelled by the duty to discover connections to drug cartels and other participants in these criminal activities, require the information to be searchable and extractable from all types of data formats. In this overview paper, we articulate an approach — a capability that uses a data description language called Data Format Description Language (DFDL) extended with higher-order functions as a host language to XML Linking (XLink) and XML Pointer (XPointer) languages in order to link, discover and extract financial data fragments from raw-data stores not co-located with each other —see figure 1. The strength of the ap-proach is grounded in the specification of a declarative compiler for our concrete language using a higher-order rewriting system with binders called Combinatory Reduction Systems Extended (CRSX). By leveraging CRSX, we anticipate formal operational semantics of our language and significant optimization of the compiler.