Rohit Jena

firstname [dot] rango [nospam at] gmail [dot] com

Iā€™m a Ph.D. student in CIS at University of Pennsylvania advised by Prof. Pratik Chaudhari and Prof. James C. Gee.

My research broadly aims to answer the following questions:

(1) How can we incorporate task-specific invariances for correspondence matching problems? (2) What kind of self-supervised representations help us discover these task-invariant representations? (3) What kind of transferability will these representations have? My research questions stem from my belief in specialist models instead of generalist ones.

I spent Summer 2024 at the NeMo team at NVIDIA where I worked on alignment of text-to-image diffusion models to improve the Pareto front of the alignment-diversity trade-off. Previously, I interned at Amazon Lab126 where I worked on mesh-NeRF hybrids for rigged 3D avatars from 360 degree videos.

I completed my Masters in Robotics at The Robotics Institute, Carnegie Mellon University where I was advised by Prof. Katia Sycara. I also worked with Prof. Kayhan Batmanghelich on segmentation for medical images. I completed my bachelors in Computer Science and Engineering from Indian Institute of Technology, Bombay in 2019. My undergraduate thesis is based on Perfect Sampling and Uncertainty Estimation in Deep Networks where I was advised by Prof. Suyash P. Awate.

Selected Publications

  1. WACV
    Elucidating optimal reward-diversity tradeoffs in text-to-image diffusion models
    Jena, R., Taghibakhshi, A., Jain, S., Shen, G., Tajbakhsh, N., and Vahdat, A.
    Winter Conference on Applications of Computer Vision 2025
    Work done at NVIDIA
    Project page PDF Supp Code
  2. NeurIPS
    Deep Learning in Medical Image Registration: Magic or Mirage?
    Jena, R., Sethi, D., Chaudhari, P., and Gee, J.
    Neural Information Processing Systems 2024
    PDF Code
  3. CVPR
    Beyond mAP: Towards better evaluation of instance segmentation
    Jena, R., Zhornyak, L., Doiphode, N., Chaudhari, P., Buch, V., Gee, J., and Shi, J.
    Conference on Computer Vision and Pattern Recognition 2023
    Highlight paper
    Acceptance rate āˆ¼2.5% of all papers, 10% of accepted papers
    Project page arXiv PDF Supp Code


Icons in the CV are adapted from Font Awesome under the Creative Commons BY license.