Date(s) - 19 Apr 2012
1:10 PM - 2:00 PM
Speaker: Robi Polikar, Chair, Rowan University Electrical and Computer Engineering
Abstract: One of the main efforts undertaken in our Signal Processing and Pattern Recognition Laboratory (SPPRL) is ensemble based system development for various computational intelligence problems that are difficult or impossible to address using single model systems. Analyzing data with missing features, predicting the confidence of an automated decision maker, incremental learning of new knowledge from streaming data without retaining or memorizing old data, fusion of heterogeneous data sources for intelligent decision making, and real world applications of these problems, such as early diagnosis of neurological disorders, phylogeny recognition in genomic data, and brain-machine interface are some of the problems SPPRL has addressed in the recent past. This talk will very briefly review these applications, and focus on our recently NSF-funded project: incremental learning of streaming data drawn from nonstationary environments, where the data characteristics – and hence underlying decision boundaries – change in time while they are being learned. This is a challenging machine learning problem, analogous to learning how to play and win a game, while the rules of the game are constantly changing without being told to the learner (player). Some applications that can benefit from such algorithms include spam detection, analysis of financial, climate, energy demand or even epidemiological data.
Bio: Robi Polikar received his B.Sc. degree in electronics and telecommunications engineering from Istanbul Technical University in 1993, M.S. and Ph.D. degrees, both co-majors in biomedical engineering and electrical engineering, from Iowa State, in 1995 and in 2000, respectively. During his stay at Iowa State, Robi worked with the Center for Non Destructive Evaluation as well as Ames Lab. In 2001 he joined Electrical and Computer Engineering at the then newly established College of Engineering of Rowan University, in Glassboro, NJ, a primarily undergraduate education program with little research enterprise and no PhD program. To demonstrate that high quality research can be done with undergraduate and MS level students, he established the Signal Processing and Pattern Recognition Laboratory at Rowan. In 2003, he received the National Science Foundation’s CAREER award – for developing incremental learning algorithms from streaming data – the first such award received by a Rowan faculty. In 2009 he received a new grant from NSF to continue his work on developing algorithms that can also learn in nonstationary environments, where the data characteristics change in time. He also restarted his MS thesis work on early diagnosis of Alzheimer’s disease, collaborating with University of Pennsylvania and receiving multimillion dollar grants first from NIH and then from Pennsylvania Dept. of Health. His current area of research interest is in ensemble based intelligent systems and their various novel applications, such as incremental learning, nonstationary learning, data fusion, and the missing feature problem in automated decision making. More recently, he also started working in brain machine interface and bioinformatics. He teaches upper level undergraduate and graduate courses in wavelet theory, pattern recognition, neural networks, signal processing, bioinformatics and biomedical systems at Rowan. In 2011, Robi was promoted to the rank of Professor and was appointed the Chair of the Electrical and Computer Engineering program. He is also tasked with developing a new Biomedical Engineering program, which Rowan intends to start in Fall 2013. Robi is a senior member of IEEE, and an Associate Editor for IEEE Transactions on Neural Networks and Learning Systems. He is currently in training to become an ABET evaluator.