Research Interests

My interest is in developing computationally efficient statistical machine learning algorithms to pattern detection in large scale data for public good. My areas of interests are in health care, digital pathology, biosurveillance, space time event detection. We use subset scanning methods to detect the subset of records that are anomalous as a group and may not be anomolous as a single record.

Current Research

My current research is focussed on detecting anomalous patterns in large scale hierarchically aggregated data. We are developing Heirarchical Linear Time Subset Scanning algorithms that detects subset of anomalous records with high accuracy in sublinear time. We are currently applying these algorithms to detect anomalous patterns in mulit-scale and multi-resolution digital pathaology whole slide images. Earlier, I have worked in biosurveillence domain in learning graph structure over which an inject might have spread and learning a graph structure to improve the detection power of spread of a disease.

Previous Research

In my previous research I worked on algorithmic game theory methods for developing bid optimizer for a search engine advertisements.