The Cocktail Party Effect - An Intro to Independent Component Analysis

Ever been chatting at a cocktail party and pick up on another conversation? If so, your brain has done some independent component analysis! That is, you have “broken a sound signal into its additive components” by maximizing the independent sub-components. This introductory talk starts with the cocktail party effect to motivate independent component analysis. We will use the R-language to demonstrate examples. We touch the history of independent component analysis and other significant applications.

About Our Speaker:

Dr. Phillip G. Bradford is a computer scientist with extensive experience in academia and industry. Phil was a post-doctoral fellow at the Max-Planck-Institut für Informatik, he earned his PhD at Indiana University, an MS form the University of Kansas, and a BA from Rutgers University. He was on the faculty at Rutgers Business School and the University of Alabama’s Engineering School. He worked for BlackRock, Reuters Analytics, founded a firm and worked with a number of early stage firms. He was a Principal Architect for General Electric and he is now on the faculty at the University of Connecticut. He is the director of the computer science program at the University of Connecticut in Stamford.

Phil has a deep belief in bringing real research to practice. This is the root of his passionate entrepreneurial perspective. Phil has a handful of best-in-class results. His Erdős Number is 2. He has given over 50 talks in nine countries and he is the author or co-author of over 65 articles.