Graduate Seminar: Combining Multiple Scoring Systems: Big Data Analytics and Cognitive Diversity
Speaker: Frank Hsu, Fordham University
Location: Torgersen Hall 2150
In this talk, I will cover methods and practices for combining multiple scoring systems (MSS’). Each of the scoring systems can exist in two different contexts: (1) as a variable (feature, parameter, cue, indicator, etc.), or (2) as a system (regression, forecasting, classification, neural net, model, decision, data mining, machine learning, etc.). In particular, we explore the issues of “when” and “how” to combine these MSS’. Conventional wisdom is that “The combination of two (or more) systems can be better than each individual system only if they are relatively good and they are diverse”. However, measurement of diversity is a challenging issue in Big Data analytics as well as in micro- and macro-informatics.
The notion of a Cognitive Diversity (CD) will be introduced. “Cognitive diversity” measures diversity between two information systems as opposed to “statistical correlation”, which measures correlation between two data distributions. CD is useful because it is simple to compute and it is independent of the data items. Based on cognitive diversity, we perform variable selection and combination (or system selection and combination). Examples are drawn from domain applications in information retrieval, target tracking, joint decision making, ChIP-seq analysis, virtual screening, cognitive neuroscience, and portfolio management.
D. Frank Hsu is the Clavius Distinguished Professor of Science, a professor of computer and information science, and director of the Laboratory of Informatics and Data Mining at Fordham University, New York, NY. He has held visiting positions at CNRS (and University of Paris-Sud), JAIST (as Komatsu Chair Professor, Kanazawa, Japan), Keio University (as IBM Chair Professor, Tokyo, Japan), MIT (Applied math and Laboratory of Computer Science), National Taiwan University, and National Tsing-Hua University (in Hsin-Chu, Taiwan).
Hsu’s current research interests include Big Data analytics, interconnection networks, machine learning, combinatorial fusion, and macro-informatics. The Combinatorial Fusion algorithm he and colleagues proposed in 2005 has been applied to diverse areas such as bioinformatics, finance, target tracking, virtual screening (and drug discovery), decision making, and cognitive neuroscience.
Hsu received a BA from National Cheng Kung University (Taiwan), an MA from the University of Texas of El Paso, and a PhD from the University of Michigan. He has served on many editorial boards including IEEE Transactions on Computers, Networks, International Journal of Foundation of Computer Science, and JOIN (Journal of Interconnection Networks). He has served as co-chairs of conference, workshops, PC’s and steering committee’s including DIMACS Workshops and I-SPAN (14th I-SPAN’ 2017 at Exeter, UK). Hsu has received the Best Paper Award in 2013 at the Brain and Health Informatics Conference in Maebashi, Japan and an IBM Faulty Award in 2012. Dr. Hsu is a Fellow of the New York Academy of Sciences and the International Institute of Cognitive Informatics and Cognitive Computing (ICIC). He is a Foundation Fellow of the Institute of Combinatorics and Application (ICA) and a Senior member of the IEEE.