News

Currently, local utilities in the U.S. generate 12 percent more energy than they anticipate needing. Otherwise, if they experience a brownout or blackout, utility companies face heavy penalties from the government. “Most of that extra energy is going to waste,” WNCG Prof. Brian Evans states. “If utility companies had access to better information, they could predict user demand and how it evolves over the day. They could reduce energy waste by around four-to-six percent.”

Prof. Joydeep Ghosh of UT ECE was the keynote speaker at the inaugural Workshop on Divergences and Divergence Learning (WDDl), held in Atlanta, June 2013. In his talk, entitled "Learning Bregman Divergences for Prediction with Generalized Linear Models," which reflects joint work with ECE and WNCG student Sreangsu Acharrya,  an efficient approach to learning a broad class of predictive models was introduced. What is most remarkable about this approach is that model parameters can be estimated even when the loss function is unknown.