(Berkeley) 2012-13 - Fall Term: CS 174: Combinatorics and Discrete Probability[course homepage]
My interests lie in machine learning and theoretical computer science. I am
primarily interested in computational learning theory -- developing efficient
algorithms for learning problems and also studying their inherent hardness. My
dissertation work was based on understanding biological evolution from a
standpoint of computational learning theory. I have also worked in on-line
learning such as experts and bandit problems.
Short Descriptions of Research Projects
Attribute-Efficient Evolvability of Sparse Linear Functions. (with E.
This short document describes our recent paper (see below) in a language
that is accessible to general scientists.
Statistical Windows in Testing for the Initial Distribution of a Reversible Markov Chain[arxiv] Quentin Berthet and Varun Kanade
In the 22nd International Conference on Artificial Intelligence and Statistics, AISTATS 2019
Learning Hurdles for Sleeping Experts[pdf][link]
Varun Kanade and Thomas
In the 3rd Innovations in (Theoretical) Computer Science Conference, ITCS
2012 Invited to a special issue of ACM Transactions on Computation Theory[submitted version]
Also presented at ICML 2012 Workshop on Exploration and Exploitation [short
Approximate Symbolic Reachability of Networks of Transition
Systems[pdf][link] Sudeep Juvekar, Ankur Taly, Varun Kanade and Supratik Chakraborty
Book chapter in Next Generation Design and Verification Methodologies for
Distributed Embedded Control Systems Springer. (Invited Paper)