Colloquium Details
Faculty Research Colloquium - Data mining and formal semantics
| Authors: | Dejing Dou University of Oregon Daniel Lowd University of Oregon |
|---|---|
| Date: | October 20, 2011 |
| Time: | 15:30 |
| Location: | 220 Deschutes |
Abstract
Presented by Daniel Lowd: Graphical models such as Bayesian and Markov networks provide compact specifications of complex probability distributions. They have been widely used in artificial intelligence applications for knowledge representation, reasoning, and learning from data. However, learning and reasoning with graphical models remains computationally difficult. In this talk, I will present recent work on improving inference in dependency networks, an alternative graphical model. In addition to sharing the results of this one project, I hope to communicate what can be done with graphical models and what conducting research in this area is like.
Presented by Dejing Dou: Traditional data mining focuses on representing discovered knowledge in a way easy for human understanding (e.g., visualization) but not necessarily for computer understanding. It is hard to reuse or share discovered knowledge with other data mining systems because of the lack of formal semantics and the heterogeneity of distributed data resources. The traditional knowledge engineering and emerging Semantic Web adopt ontologies, the formal specification of concepts and relationships, to describe semantics of data and knowledge. Ontologies have been widely used in information integration. However, there has been less research on applying ontologies in data mining in a formal way to address the challenges of knowledge sharing and reuse in a distributed and heterogeneous environment.
In the talk, Dejing Dou will introduce the research on data mining and ontologies in several projects: i) the Neural ElectroMagnetic Ontologies (NEMO) project addresses the challenge to mine and represent ontological concepts and relationships from domain specific data and to match the data tables with different semantics by using clustering and classification; ii) the Knowledge Translation project addresses the challenge to translate mined knowledge from one data resource to another semantically heterogeneous one by using logic inference; iii) the Semantic Association Mining project addresses the challenge to mine indirect associated items which are annotated with medical ontological concepts by using Hypergraphs. Our research contributes to a larger theme of semantic data mining, in which formal semantics (e.g., ontologies) and semantic linkages that exist in data can be discovered and incorporated into the knowledge discovery and knowledge sharing processes.
Biography
Dejing Dou is an Associate Professor in the Computer and Information Science Department at the University of Oregon and leads the Advanced Integration and Mining (AIM) Lab. He received his bachelor degree from Tsinghua University, China in 1996 and his Ph.D. degree from Yale University in 2004. His research areas include ontologies, data integration, data mining, biomedical and health informatics, and the Semantic Web. He has published more than 40 research papers, some of which appear in prestigious conferences and journals like KDD, ICDM, SDM, CIKM, ISWC, ODBASE, JIIS and JoDS. His KDD'07 paper was nominated for the best research paper award. In addition to serving on numerous program committees, Dejing Dou has been invited as panelist by the NSF several times, and as an expert for grant review by the Netherlands Organization for Scientific Research (NWO). He is on the Editorial Board of Journal of Data Semantics. Dejing Dou has received over $3 million PI or co-PI research grants from the NSF and the NIH.
Daniel Lowd is an Assistant Professor in the Department of Computer and Information Science at the University of Oregon. His research covers a range of topics in statistical machine learning, including statistical relational representations, unifying learning and inference, and adversarial machine learning applications (e.g., spam filtering). In 2009, he coauthored book on Markov logic with Pedro Domingos, published by Morgan & Claypool. He is also the recipient of graduate research fellowships from the National Science Foundation and Microsoft Research.