Doug Burdick has been a Research Staff Member at IBM Research-Almaden in San Jose, Calif., since 2010.  His research focuses in information extraction and entity integration, with application to the financial domain.  Burdick was also an early contributor to Apache SystemML, which is an open source scalable machine learning platform originally developed at IBM Research.  Prior to joining IBM Research, Burdick was a researcher at MITRE from 2007 to 2010 and focused on schema matching and cyber security.  He received his PhD in Computer Science from the University of Wisconsin-Madison in 2007, and his BS and MEng in Computer Science from Cornell University in 2000 and 2001.

Abstract

Recent development of Big Data toolkits has enabled creation of high-quality datasets to support both system-wide and granular modeling of the financial system. Such datasets enable researchers to create novel modeling techniques, and provide financial regulators deeper insights into potential issues.

This talk will provide a brief overview of joint efforts with University of Maryland, Office of Financial Research, NIST, and IBM to build a community of academic researchers, industry participants, and financial regulators for developing toolkits to construct such datasets and novel modeling techniques to fully leverage these datasets.  Concrete results of this collaboration include the Data Science for Macro-Modeling (DSMM) workshops (http://www.dsmmworkship.org), and the Financial Entity Identification and Information Integration (FEIII) Challenge (https://ir.nist.gov/dsfin/).