I defended my Ph.D in AI from the Cheriton School of Computer Science at the University of Waterloo.
I am interested in the theory and practice of machine learning. My thesis was focussed on developing and improving the theoretical foundations of clustering.
I am fortunate to be advised by Prof. Shai Ben-David. A consequence of this is that my research statement closely matches his..
When we discuss with some practioners as to why they chose a particular clustering algorithm, we often get the following responses. "Everyone uses it". OR "It has a nice python/matlab/(pick your favorite language) implementation". OR "Runs fast".
My research aims to address these issues by developing a principled way of choosing clustering algorithms. When is clustering easy? On what datasets should a particular be used? How should we incorporate human expertise into the clustering problem? We provide answers to such questions.
- Solve the problem of detecting duplicates in databases using principled clustering techniques.
- Framework of incorporating domain knowledge into the clustering problem.
Finding meaningful cluster structure amidst background noise.[ALT' 16]
Multi-Pivot Quicksort: Theory and Experiments.[ALENEX' 14]
- Provably fastest version of quicksort.