Dr. Fisher's research emphasizes machine learning and information theory applied to distributed sensing and inference under resource constraints. A primary goal is to develop computationally efficient algorithms with provable theoretical performance guarantees. Dr. Fisher received his Ph.D. degree in Electrical and Computer Engineering from the University of Florida, Gainesville, Florida in 1997.

Decision making is a challenging process. Uncertainty in the outcome may be difficult to quantify. Information sources may be numerous and complex. Time is often a critical limitation. As a result, individuals and organizations expend significant resources -- reasoning, information gathering, energy, and time -- as part of the decision-making process. While recent advances demonstrate the potential for machine learning methods to facilitate decision-making, many methods rely on extensive manually labeled datasets for training. Such datasets may not be easily obtained for a given problem. In his talk, Dr. Fisher will highlight some of the appealing properties of structured probabilistic models for integrating disparate information sources in a manner that exploits recent advances in machine learning without requiring explicitly labeled training data. These representations provide not only "answers", but also enable Value of information (VoI) analysis to explore the trade-off between the informational utility of a distributed set of information sources, the resources necessary to acquire them, and their impact on decision risk. Dr. Fisher will present a variety of real-world applications of these methods highlighting both their advantages as well as pointing to new challenges.