Fares Hedayati

 

Email: fareshed at eecs.[university].edu

 

I obtained my Ph.D. from UC Berkeley, working with Prof. Peter Bartlett. My primary research interests are in theoretical machine learning, online learning and minimum description length principle. I have also worked on OCR (Optical Character Recognition) and handwriting recognition. Here is a link to my CV on linkedin.

 

Publications

 

Ph.D. Dissertation

Minimax Optimality in Online Learning under Logarithmic Loss with Parametric Constant Experts [pdf]

 

 

The Optimality of Jeffreys Prior for Online Density Estimation and the Asymptotic Normality of Maximum Likelihood Estimators [pdf]

Fares Hedayati, Peter Bartlett

 

Exchangeability Characterizes Optimality of Sequential Normalized Maximum Likelihood and Bayesian Prediction with Jereys Prior [pdf]

Fares Hedayati, Peter Bartlett

 

Recognition of Tibetan wood block prints with generalized hidden Markov and kernelized modified quadratic distance function [pdf]

Fares Hedayati, Jike Chong, Kurt Keutzer

 

Horizon-Independent Optimal Prediction with Log-Loss in Exponential Families [pdf]

 Peter Bartlett, Peter Grunwald, Peter Harremoes, Fares Hedayati, Wojciech Kotlowski



Teachings

 

 

CS70: Discrete Mathematics and Probability Theory

Spring 2011, James Demel, UC Berkeley 

 

CS170: Efficient Algorithms and Intractable Problems

Fall 2008, Satish Rao, UC Berkeley 

 

CS61A: Structure and Interpretation of Computer Programs 

Fall 2009, Brian Harvey, UC Berkeley