NSE - Nuclear Science & Engineering at MIT

FAQ | Contact | Jobs | NSE Policies

Machine Learning Research and Applications at IBM

Monday, November 18th, 4:00 to 6:00 PM
MIT Stata, 32-123
Reception to follow

Co-hosted by NSE, EECS GSA and MIT Sloan School of Management.

Dr. Rick Lawrence
Machine Learning Group
IBM Research
Title: An SMS Text Classification System for UNICEF Uganda

Dr. Dmitry Malioutov
Machine Learning Group
IBM Research
Title: Sparse Learning in Finance

Dr. Aurélie Lozano
Machine Learning Group
IBM Research
Title: Minimum Distance Estimation for Robust and Sparse High-Dimensional Regression

Abstracts & Bios

An SMS Text Classification System for UNICEF Uganda

Dr. Rick Lawrence
Machine Learning Group
IBM Research
Yorktown Heights, NY

U-report is an open-source SMS platform operated by UNICEF Uganda, designed to give community members a voice on issues that impact them. Data received by the system are either SMS responses to a poll conducted by UNICEF or unsolicited reports of problems occurring anywhere within Uganda. There are currently over 240,000 U-report participants, and they send up to 10,000 unsolicited text messages a week. The objective of the program in Uganda is to understand the data in real-time, and to insure that critical issues identified in the messages are addressed by the appropriate department in UNICEF in a timely manner. This talk describes an automated message-understanding and routing system deployed by IBM Research at UNICEF. We discuss a dual-supervision learning technique to leverage human-generated labels on both features and text examples, and conclude with a discussion of the societal impact that U-report is already driving in Uganda.
Joint work with Prem Melville, Vijil Chenthamarakshan, and Solomon Assefa (IBM Research)

Rick Lawrence manages the Machine Learning Group in IBM Research at Yorktown Heights NY. He led the IBM sales productivity project selected as a Finalist in 2009 for the prestigious INFORMS Franz Edelman competition that recognizes the application of advanced analytics to solve challenging business problems. He currently leads several research projects involving the development and application of machine learning methods to extract insight from social media and other sources of unstructured content. One current project, with UNICEF in Uganda, received the Best Paper Award in the Industry & Government Track at the 2013 ACM KDD Conference, and is a Finalist in the 2014 INFORMS Innovative Applications in Analytics Award. Rick received his undergraduate degree at Stanford in Chemical Engineering and his Ph.D. in Nuclear Engineering at the University of Illinois.


Sparse Learning in Finance

Dr. Dmitry Malioutov
Machine Learning Group
IBM Research
Yorktown Heights, NY

Statistical modeling in high dimensions has lately become one of the most active research areas in statistics, machine learning and signal processing, introducing a variety of exciting topics such as sparse approximations, nuclear norm relaxations, graphical model structure learning and compressed sensing, among many others. In this talk we overview recent applications of sparse learning to finance that have resulted in new approaches for portfolio selection, hedging, statistical arbitrage, and factor modeling.

Dmitry Malioutov is a research staff member in the Machine Learning group in the Business Analytics and Mathematical Sciences (BAMS) department at the IBM T.J. Watson Research Center, Yorktown Heights, NY. Prior to joining IBM, Dmitry had spent several years as an applied researcher in high-frequency trading in DRW Trading, Chicago, and as a postdoctoral researcher in Microsoft Research, UK. Dmitry received the Ph.D. and the S.M. degrees in EECS from MIT.

His research interests include inference and learning in graphical models, message passing algorithms; Sparse signal representation; Sensor array source localization; Statistical risk modeling, robust covariance and joint dependence modeling; portfolio optimization. More generally: statistical signal processing, machine learning and convex optimization; applications in quantitative finance.


Minimum Distance Estimation for Robust and Sparse High-Dimensional Regression

Dr. Aurélie Lozano
Machine Learning Group
IBM Research
Yorktown Heights, NY

We propose a minimum distance estimation method for robust regression in sparse high-dimensional settings. The traditional likelihood-based estimators lack resilience against outliers, a critical issue when dealing with high-dimensional noisy data. Our method, Minimum Distance Lasso (MD-Lasso), combines minimum distance functionals, customarily used in nonparametric estimation for their robustness, with l1-regularization for high-dimensional regression. The geometry of MD-Lasso is key to its consistency and robustness. The estimator is governed by a scaling parameter that caps the influence of outliers: the loss per observation is locally convex and close to quadratic for small squared residuals, and flattens for squared residuals larger than the scaling parameter. As the parameter approaches infinity, the estimator becomes equivalent to least-squares Lasso. MD-Lasso enjoys fast convergence rates under mild conditions on the model error distribution, which hold for any of the solutions in a convexity region around the true parameter and in certain cases for every solution. Remarkably, a first-order optimization method is able to produce iterates very close to the consistent solutions, with geometric convergence and regardless of the initialization. A connection is established with re-weighted least-squares that intuitively explains MD-Lasso robustness. The merits of our method are demonstrated through simulation and eQTL data analysis.
Joint work with Nicolai Meinshausen, Professor of Statistics, University of Oxford and ETH Zürich.

Aurélie Lozano received her Ph.D. from Princeton University, where she was a recipient of the Gordon Y.S. Wu Fellowship. Since 2007, she has been a research staff member in the Machine Learning group at the IBM T.J. Watson Research Center, Yorktown Heights, NY. Dr. Lozano's research interests include machine learning, statistics and data mining. Her current focus is on methods for solving high dimensional data problems, their theoretical analysis, and applications to computational biology, environmental sciences, and social media analytics.

RicK Lawrence

Dr. Rick Lawrence
Machine Learning Group
IBM Research
Yorktown Heights, NY

Dmitry Malioutov

Dr. Dmitry Malioutov
Machine Learning Group
IBM Research
Yorktown Heights, NY

Aurélie Lozano

Dr. Aurélie Lozano
Machine Learning Group
IBM Research
Yorktown Heights, NY

Massachusetts Institute of Technology

Facebook

Copyright © 2014 Department of Nuclear Science and Engineering