CURRENT RESEARCH PROJECTS
Semantic Mining of Activity, Social, and Health data (SMASH) is a multiple-discipline research among computer scientists, medical doctors, and social scientists. Traditionally, support groups and other social reinforcement approaches have been popular and effective in dealing with unhealthy behaviors including overweight. Research in the design and implementation of the SMASH (Semantic Mining of Activity, Social, and Health data) system will address a critical need for formal ontologies and data mining tools to help understanding the influence of healthcare social networks, such as YesiWell, on sustained weight loss where the data are multi-dimensional, temporal, semantically heterogeneous, and very sensitive. SMASH will develop new methods in social network analysis, Semantic Web ontologies, privacy preserving data mining, building on the current state-of-the-art. This project is being supported with a three-year R01 grant by the NIH/NIGMS (Grant Number: R01GM103309 $1.54M PI: Dejing Dou; Co-I: Brigitte Piniewski, Ruoming Jin, Xintao Wu, Jessica Greene, Daniel Lowd, Junfeng Sun; Consultant: David Kil; 5/1/2013 - 2/29/2016).
Statistical Knowledge Translation and Integration (SKTI)
combines formal ontologies and Markov logic to thoroughly address the challenging problem of translating and integrating semantically heterogeneous knowledge in a systematic way. By expressing both knowledge and semantic mappings in formal ontologies and Markov logic, a unified probabilistic model can jointly translate and integrate knowledge with uncertain mappings. The methods will be evaluated in real-world
ontologies, knowledge bases, and a benchmarking system for heterogeneous data. This research will contribute to distributed data mining, knowledge transfer, and 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 process. This project is being funded with a three-year research grant by the NSF (Award Number:
IIS-1118050 $495K PI: Dejing Dou Co-PI: Daniel Lowd 7/1/2011 - 6/30/2014).
Neural ElectroMagnetic Ontologies (NEMO)
need for formal representation, storage, mining, and dissemination of brain
electromagnetic (e.g., EEG) data. NEMO is a collaborative project between computer scientists
and neuroscientists. NEMO aims to develop ontologies and ontology-based methods for representing and sharing event-related brain potentials (ERP) data from experimental studies of neural processes underlying human language and cognition. The NEMO project is the first to develop formal ontologies for the ERP domain. These ontologies are used to represent the current state of knowledge in the ERP domain and to support ontology-based mark-up (annotation) of ERP experiment data collected within our NEMO consortium. With the ontology-based mining, mapping, and integration tools developed for this project, the NEMO research team aims to conduct meta-analyses of ERP patterns in language and cognition, combining results from a variety of ERP research paradigms and different analysis methods and results from our international team of ERP researchers. The NEMO project is being funded with a four-year R01 grant by the NIH/NIBIB (Grant Number: R01EB007684 $2.22M PI: Dejing Dou Co-I: Gwen Frishkoff, Allen Malony, Don
Tucker 5/1/2009 - 4/30/2013).
Ontology Based Information Extraction (OBIE) has recently emerged as a sub-field of information extraction (IE). Here, the general idea is to use ontologies to guide the information extraction process and formally present the results of information extraction. We have focused on three directions: First, identifying components of information extraction systems that make extractions with respect to particular components of an ontology (which we call information extractors) and reuse those information extractors in other IE processes. Second, using multiple ontologies in the same domain with their semantic mappings to improve the performance of IE. Third, constructing and enriching ontologies automatically during the IE processes.
PREVIOUS RESEARCH PROJECTS
OntoGrate is an ontology-based information integration
framework. The general goal is to integrate information that is heterogenous
in both structure and semantics in a highly automatic way. Key innovations
in OntoGrate include broadening the typical scope of integration to span
databases, XML data, and the Semantic Web; strengthening and formalizing the
derivation of mapping rules by introducing machine learning and data mining
techniques; and extending our inference engine, OntoEngine, and first order
ontology language, Web-PDDL, to solve the problem of integration using
formal mapping rules with uncertainty. It is novel to apply multi-relational
data mining to discover complex mapping rules. As one application of
OntoGrate, we have collaborated with the
ZFIN (Zebrafish Model
Organism Database) research group to integrate heterogeneous gene
Integration). The OntoGrate project was supported by the start-up fund of
Dejing Dou from the
University of Oregon.
Internet Routing Forensics (IRF) is a collaborative
project with Jun Li at the UO's
Network Security Lab
and David Meyer at the
Advanced Network Technology Center. We are extending several data mining
techniques, such as classification and clustering, to discover and analyze
abnormal BGP (Border Gateway Protocol) events, such as worms and blackouts.
This project was funded with a three-year research grant by the
NSF (Award Number:
CNS-0520326 $350K PI: Jun Li Co-PI: Dejing Dou, David Meyer 10/1/2005 - 9/30/2008).