Title
Understanding the computational models for dense correspondence matching
Speaker Bio
Rohit Jena is a Ph.D. candidate in Computer and Information Science at the University of Pennsylvania, where he is co-advised by Professors James C. Gee and Pratik Chaudhari. His research focuses on developing and unraveling the nature of task-specific, self-supervised representations for correspondence matching problems, with an emphasis on specialist models over generalist ones.
Previously, Rohit did his Master’s in Robotics from Carnegie Mellon University’s Robotics Institute where he worked on enabling autonomous AI to understand mental models of human experts to enable human-machine collaboration. He earned his Bachelor’s in Computer Science and Engineering from the Indian Institute of Technology, Bombay, completing a thesis on perfect sampling and uncertainty estimation in deep networks.
Rohit’s work has been recognized at premier conferences, including CVPR, NeurIPS, MICCAI, WACV and SPIE where he has presented multiple oral talks. He has also reviewed more than 50 papers at multiple machine learning and AI conferences.
Rohit has also worked at leading research labs including NVIDIA’s NeMo team, where he explored reward-diversity tradeoffs in text-to-image diffusion models, and Amazon Lab126, where he worked on mesh-NeRF hybrids for real-time human avatar generation. He also worked on AI for edge computing at Microsoft.
Schedule:
Date: Friday, August 15th 2025
Time: 09:00 PST | 17:00 West Africa Time | 16:00 UTC
Zoom Meeting Link: https://us06web.zoom.us/j/89709172931
Contact Information
compass@mailab.io
View recordings of previous sessions on our YouTube channel https://www.youtube.com/@COMPASS-nn5or
Leave a Reply