December 27, 2019: Seminar by Dr. Prithviraj Dasgupta, U.S. Naval Research Laboratory, Washington D.C.

An Overview of Adversarial Training Techniques for Countering Adversarial Attacks in Machine Learning
by
Dr. Prithviraj Dasgupta
Distributed Intelligent Systems Section (Code 5583)
Information Management and Decision Architectures Branch
Information Technology Division
U.S. Naval Research Laboratory
4555 Overlook Avenue SW
Washington D.C. 20375
 
 
Date: Friday, December 27, 2019
Time: 3:00-4:00 PM
Venue: Seminar Room, 
            Electronics and Communication Sciences Unit 
            9th Floor, S. N. Bose Bhavan (Library Building)
 
 
Abstract
Deep learning techniques are known to be vulnerable to adversarial attacks including evasion and data poisoning attacks, during both training and testing times. Several techniques, called adversarial training, have been proposed to build defenses of the model learned by a machine learning algorithm against adversarial attacks. This talk will provide an overview of the major adversarial training techniques, their vulnerabilities and limitations, and ongoing research directions in this area.
 
Bio: Dr. Prithviraj (Raj) Dasgupta is a research scientist at the U.S. Naval Research Laboratory, Washington D.C. His areas of interest  include multi-agent systems, game theory and machine learning. From 2001 through 2019, he was a professor in the Computer Science department at the University of Nebraska, Omaha, where he had founded and directed the CMANTIC Robotics Lab. He has published over 150 papers in leading journals and conferences in his area and received the highest research award from the University of Nebraska, Omaha,  called ADROCA, in 2017. He received his Ph.D. and M.S. in Computer Engineering from the University of California, Santa Barbara and his B. Engg. in Computer Science from Jadavpur University.
 
 
All are cordially invited.
 
 
Dipti Prasad Mukherjee
Head, Electronics and Communication Sciences Unit