ENEE729T

Prerequisite: ENEE 620 or equivalent; and elements of information theory This course covers several information theoretic methods of interest in statistical inference and machine learning. Topics include: (i) information geometry leading to the EM algorithm and applications; (ii) measure concentration methods; (iii) correlated multiarmed bandits including in probability distribution learning from partially sampled observations (finding an arm or a small set of arms that yield information about other correlated arms) with applications in sensor placement in IoT or for environmental or cellular network monitoring; and (iv) applications in data privacy vs function computation utility tradeoffs.