Uddeshya Upadhyay
PhD Researcher
Github Website uddeshya.upa at gmail.com

Profile

I am a first-year Ph.D. candidate at IMPRS-IS Program, Tübingen, where I am part of EML and MIDAS group, and I work closely with Prof. Zeynep Akata and Prof. Sergios Gatidis. I am fascinated by interdisciplinary R&D happening at the intersection of Computer Vision, Machine Learning, Biomedical and Healthcare Informatics. These days I am exploring Bayesian Deep Learning, Uncertainty Estimation, Generative models, and Explainable AI. I was an undergrad at Computer Science and Engineering@IIT-Bombay.


Earlier, I spent some time as a research intern at Microsoft Research, Honda Research Institute, and Nanyang Technological University-Singapore. Besides, I've also spent time working with medical imaging startups and as a quantitative researcher for algorithmic trading firm.


For more details about me, please check my homepage.

Publications

  • "BayesCap: Bayesian Identity Cap for Calibrated Uncertainty in Frozen Neural - Networks", Uddeshya Upadhyay*, Shyamgopal Karthik*, Yanbei Chen, Massimiliano Mancini, Zeynep Akata, European Conference on Computer Vision, ECCV 2022
  • "Robustness via Uncertainty-aware Cycle Consistency", Uddeshya Upadhyay, Yanbei Chen, Zeynep Akata, Neural Information Processing Systems, NeurIPS 2021
  • "Uncertainty-Guided Progressive GANs for Medical Image Translation", Uddeshya Upadhyay, Yanbei Chen, Tobias Hepp, Sergios Gatidis, Zeynep Akata, Medical Image Computing and Computer Assisted Interventions, MICCAI 2021