RESEARCH   

Artificial Intelligence   
Biomorphic Robotics Lab   
Computational Biology & Bioinformatics   
Computer Architecture & VLSI   
Computer Vision and Pattern Recognition   
Distributed Systems   
Information Systems   
Robotics   
Software Security   
REU Program   
Seminars   

Research::Artificial Intelligence   

INTELLIGENT SYSTEM LABORATORY
http://isl.cse.usf.edu/ailab/aipage

Sasha Dos-Santos, Todd Winchell, and Sean Barbeau

The Intelligent Systems Laboratory faculty in the Computer Science and Engineering department involves Ph.D., M.S., and undergraduate students in research projects spanning many areas of Artificial Intelligence. There is close collaboration with other department faculty from the Vision and Network research groups as well as with faculty from the College's Center for Urban Transportation Research. On-going research projects include the scaling of machine learning algorithms to very large data sets, the use of expert systems with GPS-enabled PDAs and cellular telephones, distributed machine learning, data mining, pattern recognition, and fuzzy logic. Research is funded by the U.S. Department of Transportation, the Florida Department of Transportation, the National Center for Transit Research, the National Science Foundation, the National Institutes of Health, and the Department of Defense.   

RECENT AND CURRENT PROJECTS

AVATAR: Large-Scale Distributed Learning   

LARGE-SCALE DATA MINING
Large-Scale Data Mining involves building models from a large number of labeled or unlabeled examples. Typically, the number of examples is too large to fit in a single main memory. This makes the model building problem intractable unless subsampling or distributed processing is used. Our approach is to build ensembles of classifiers for labeled data in a distributed fashion. We have developed both distributed boosting algorithms and fully distributed learning algorithms. These have been shown to be highly effective, even for data sets which will fit within a single main memory. With unlabeled data, we have investigated a distributed clustering scheme where the results are merged to provide an overall partition for very large-scale data. This approach has shown significant promise. This work can be applied to any large repository of data to build both predictive and descriptive models.

TRAC-IT: Expert System for Travel Advice   

PERSONAL DIGITAL TRAVEL DIARY
A Personal Digital Travel Diary is being developed in collaboration with the Center for Urban Transportation Research. This wireless communication device uses GPS-enabled PDAs and cellular telephones to track a person's travel route as well as the mode (walking, bicycling, car, or bus) of travel. This device will provide more reliable data for the Department of Transportation to justify construction of new highways and roads and their repair and improvement. Real-time communication will provide the users of this device alternate routes based on real-time incident and traffic conditions and, if appropriate, feedback suggesting improvements in travel behavior.