Omprakash Chakraborty

Postdoctoral Fellow at École de technologie supérieure (ÉTS), Montréal

FRQnet Postdoctoral Fellow (2025) Google PhD Fellow (2023)

I work in computer vision focusing on areas of vision–language models, label-efficient learning, test-time adaptation, along with calibration and reliable deployment in medical imaging.

omprakash.chakraborty@livia.etsmtl.ca · omchakrabarty@gmail.com · Google Scholar

Omprakash Chakraborty

News

2026

  • June:
    Paper accepted at ECCV 2026QaTS: Quantile-Adaptive Temperature Scaling for Confidence Calibration.
    Paper accepted at MICCAI 2026IMaX: Information Maximization for Long-Tailed Semi-Supervised Domain Generalization.
  • Feb: Two papers accepted at CVPR 2026ORION: ORthonormal Text Encoding for Universal VLM AdaptatION and Semantic Orthogonal Calibration (SoC).

2024

  • Nov: TMLR paper accepted — XPL: Semi-Supervised Prompt Learning for VLMs.
  • Jun: Received Nasscom AI Gamechangers recognition.

2023

  • May: ICLR paper — AnyDA: Anytime Domain Adaptation.
  • Apr: Awarded Google PhD Fellowship.
  • Jan: Received INSC Young Researcher Award.

Selected Publications

O. Chakraborty, J. Dolz, I. Ben Ayed
ORION: ORthonormal Text Encoding for Universal VLM AdaptatION.
CVPR, 2026.

L. Fillioux, O. Chakraborty, I. Ben Ayed, P.H. Cournède, S. Christodoulidis, M. Vakalopoulou, J. Dolz.
SoC: Semantic Orthogonal Calibration for Test-Time Prompt Tuning.
CVPR, 2026.

O. Chakraborty, L. Fillioux, I. Ben Ayed, J. Dolz.
Quantile‑Adaptive Temperature Scaling for Confidence Calibration.
ECCV, 2026.

L. Fillioux, O. Chakraborty, I. Ben Ayed, P.H. Cournède, S. Christodoulidis, M. Vakalopoulou, J. Dolz.
Information Maximization for Long-Tailed Semi-Supervised Domain Generalization.
MICCAI, 2026.

O. Chakraborty, A. Sahoo, R. Panda, A. Das.
XPL: A Cross-Model Framework for Semi-Supervised Prompt Learning in Vision-Language Models.
Transactions on Machine Learning Research (TMLR), 2024.

O. Chakraborty, A. Sahoo, R. Panda, A. Das.
AnyDA: Anytime Domain Adaptation.
International Conference on Learning Representations (ICLR), 2023.

O. Chakraborty, A. Singh, A. Varshney, R. Panda, R. Feris, K. Saenko, A. Das.
Semi-Supervised Action Recognition with Temporal Contrastive Learning.
CVPR, 2021.

Research Areas

Label-efficient Learning

Efficient learning with limited labels and compute; semi-self supervised, prompt-based methods.

Vision–Language Models

Parameter-efficient adaptation; robust prompts; medical imaging transfer.

Medical Imaging (Ophthalmology)

Exploring the field of ophthalmology to handle tasks such as calibration and interpretability

Anytime / Efficient Inference

Anytime inference, dynamic computation, and resource-aware recognition.

CV

Download CV (PDF)

Contact

Email: omprakash.chakraborty@livia.etsmtl.ca | omchakrabarty@gmail.com

Profiles: GitHub · LinkedIn · Facebook