Speaker : Prof. Fushing Hsieh (University of California)
Title : 1.Data-driven computing and graphic display. 2.Pursuit intelligence from data or algorithm?
Time : 2018-03-05 (Mon) 14:30 - 17:00
Place : Seminar Room 722, Institute of Mathematics (NTU Campus)
Abstract: 1.I begin by asking a simple question: what role of a computing scientist in working on a big database? I illustrate my thinking trajectory along with data-driven computing on PITCHf/x database: first, for extracting information of MLB pitcher’s pitching mechanics; secondly, for representing computed information via graphic displays; thirdly, for and stimulating understanding and knowledge. At the end, I mention the merits of the three steps. The PITCHf/x database is available in Major League Baseball Advanced Media(MLBAM) website.
2.I start with questioning: whether we are able to harvest intelligence from black-box algorithm, such as Random Forest or Deep Learning? Then I turn to address a futuristic question: what if we can indeed discover and harvest intelligence by data-driven computing? For both questions, I argue and discuss based on my experimental results on various data sets.