June 24, 2009Smyth Awarded 2009 SIGKDD Innovation Award
Padhraic Smyth, Professor of Computer Science at the Bren School of Information and Computer Sciences, has been awarded the 2009 Innovation Award from the Association of Computing Machinery’s Special Interest Group on Knowledge Discovery and Data Mining (SIGKDD). This award is given annually to an individual who has made outstanding research and technical contributions to the field of data mining.
Smyth, also Director of the Center for Machine Learning and Intelligent Systems at UC Irvine, is being recognized for his contributions to both the theory and application of probabilistic and statistical approaches to data mining. He will receive the award at the 15th annual ACM SIGKDD Conference to be held June 28 – July 1 in Paris, France. Smyth’s research is in the area of statistical data mining and machine learning. His research focuses on both the basic principles of inference from data (theory and algorithms), combined with applications to a variety of data-driven problems in the sciences, medicine, and engineering.
A central theme in Smyth’s work is the use of probabilistic modeling of complex and large data sets using hidden variable methods. An example of how his research is applied in practice is his work in climate and planetary science, including clustering of storm tracks in the Atlantic and Pacific oceans, seasonal forecasting of rainfall in the tropics, analysis of global geopotential height patterns, and classification algorithms for detecting volcanoes in images of Venus.
Smyth is the recipient of a $250,000 grant funded the US Department of Energy's Climate Change Prediction program as part of the SciDAC initiative (Scientific Discovery Through Advanced Computing).
As part of this research, along with colleagues from UCLA, Columbia University and University of Wisconsin, Smyth is investigating techniques for the development of large-scale statistical time-series models for forecasting of seasonal rainfall patterns in regions such as India.
Making such forecasts more accurate is important from a variety of economic and societal viewpoints. Statistical machine learning algorithms provide a very useful way to automatically analyze historical data to build better predictive models.
Other applications of Smyth’s research include prediction and analysis of patterns over time in social networks, clustering of user navigation patterns on Web sites, spatial modeling of brain images, analysis of time-course gene expression data, and event detection in large-scale traffic sensor data.
Smyth received his MSEE and Ph.D. degrees from the California Institute of Technology in 1985 and 1988 respectively. He has served as an associate editor for the Journal of the American Statistical Association, the IEEE Transactions on Knowledge and Data Engineering, and the Machine Learning Journal, and has served as an editorial board member for the Journal of Data Mining and Knowledge Discovery and the Journal of Machine Learning Research.
