Knowledge Enhanced Data Mining and Machine Learning

Speaker: Prof. Ian Davidson
Date: Friday, September 9, 2011
Time: 11:15am-12:30pm
Location: 2150 Torgerson

 

Abstract:
Traditional algorithms in data mining and machine learning cannot easily encode a domain expert's knowledge, hence the chance of them finding novel or complimentary insights is limited. In this presentation I will sketch some results for encoding domain knowledge into new and existing algorithms in a principled manner. We show how to encode domain expertise into a number of problems including graph clustering/segmenting and hierarchy building and will present application results for fMRI and text analysis.

Bio:
Ian Davidson is an associate professor of computer science at the university of California - Davis where he directs the knowledge discovery and data mining lab. His work focuses on principled data mining algorithm development in particular encoding knowledge into algorithms. With collaborators he has applied his work to a number of different domains including social network analysis in Afghanistan (with MITRE), war game simulations (with U.S. Army) and fMRI analysis (Alzheimer's institute at U.C. Davis). He received his Ph.D. in 2000 from Monash University in Australia. He has been the recipient of best paper awards at the SIAM and IEEE data mining awards and is on the editorial board of most data mining journals. His research is supported by ONR, NSF, OSD, Google and IBM awards.