Despite its importance in many applications, anomaly detection using one-class classification techniques, is notorious for its inability to scale to large features and datasets. We offer FROCC family of one-class classification techniques that are super-fast (12x fast) and accurate (2 pp better than closest deep learning based method). Oh…it scales to millions of dimensions and training datasets easily!
I am an associate professor and DS Chair Professor of Artificial Intelligence at the Department of Computer Science and Engineering at IIT Delhi. I am also closely associated with the Amar Nath and Shashi Khosla School of Information Technology and the newly formed ScAI - School of Artificial Intelligence.
I have previously worked as a research scientist at IBM Research as part of their AI Research team, as a faculty member at IIIT Delhi, and senior researcher at the Max-Planck Institute for Informatics (MPII) with Gerhard Weikum. My PhD (2005) was from the Indian Institute of Science where I worked with Jayant Haritsa.
Interests
- Large-scale graph data management
- Influence Propagation and Community Formation on Location- and Activity-based Social Networks
- Knowledge-bases - infrastructure, modelling and practical usage
- Data wrangling - data integration, relationship discovery and visualization
- Information Retrieval using KB and neural models
Education
-
PhD, 1998-2005
IISc
-
BE, 1991-1995
Bangalore University
