I am a third year Ph.D. student in Machine Learning Department in Carnegie Mellon University, supervised by Dr. Jeff Schneider. I like methods that are simple but not superficial, effective but not sophisticated, generative but not with degenerated performance.
I am interested in learning processes with human/new data in the loop. Examples are active/semi-supervised learning, outlier detection, and reinforcement learning. The challenges here are that assumptions need to be valid yet generalizable, that heuristics need to bear theoretical explanations, and that optimization has to be done fast.
Active Search and Bandits on Graphs Using Sigma-Optimality
A second paper in the series of Sigma-optimality, a mysterious idea which seems crazy at first but actually provokes thoughts.
Active Pointillistic Pattern Search
Active Pointillistic Pattern Search, paper and code
Active Area Search via Bayesian Quadrature
Gateway to "AAS via BQ". Paper and Code
Sigma-Optimality in Active Learning on Gaussian Random Fields
A new heuristic proposed. Algorithmic bounds discovered.