Abstract of the Presentation
- 1. American University of Armenia COLLEGE OF ENGINEERING SEMINAR SERIES Statistical Machine Learning and Data Mining: A Brief Survey by Aram Galstyan, Research Scientist USC Information Sciences Institute 4676 Admiralty Way, Marina del Rey, CA 90292, USA http://www.isi.edu/~galstyan Abstract of the Presentation Machine Learning (ML) has been the Holy Grail of the AI ever since its inception in the late 50-s. Unfortunately, the efforts to build machines with human-like intelligence in a general sense of the word have been largely unsuccessful. Instead, recent research, helped by a surge in the amount of available training data, has focused on more data driven and statistical approaches to learning. This more pragmatic type of ML has been tremendously successful in developing intelligent systems and technologies for automated speech recognition, machine translation, image processing, and so on. In this talk I'll survey the current ML research for data mining, which is concerned with detecting regularities and patterns in data, be it a collection of texts or web-pages, genetic sequences, or time series describing various processes. Specifically, I'll give a brief and self-sufficient introduction to the several forms of statistical learning, elaborate on unsupervised, supervised, and semi-supervised learning, and provide a few examples of successful applications of current ML techniques. I will suggest that, a side from pure academical interest, this type of research has a huge practical and economical potential. A desired by-product of this talk is to identify potential thesis topics for interested students. About the Speaker Aram Galstyan is a research scientist at the USC Information Sciences Institute, Intelligent Systems Division. His current research mainly focuses on learning and discovering patterns in large scale spatial- temporal data. His other research interests include multi-agent systems, social network analysis, and more generally, use of statistical physics for modeling and analyzing complex adaptive systems. Dr. Galstyan holds an MS in Physics from the Yerevan State University, and a PhD in theoretical condensed matter physics from the University of Utah. He is a member of the American Association of Artificial Intelligence (AAAI) and has authored more than 30 technical papers on various topics in computer science and physics. Date: Thursday, October 25, 2007 Time: 16:00-17:30 Place: Small Auditorium, 5th floor AUA The seminar is open to the public.