Digbalay Bose

Digbalay_Bose.png

I am a Research Scientist at Adobe Research, Bengaluru, where I currently work on improving multi-modal document understanding. I obtained my Ph.D. in Electrical Engineering from the Department of Electrical and Computer Engineering at the University of Southern California, Viterbi School of Engineering. I was advised by Prof. Shrikanth Narayanan at the Signal Analysis and Interpretation Laboratory. Prior to USC, I worked as a Research Software Engineer at IBM Research, India from 2016 to 2018. I received my Masters in Electrical Engineering from Indian Institute of Technology, Bombay in 2016, where I was advised by Prof. Subhasis Chaudhuri. Previously, I obtained my Bachelors degree in Electronics and Telecommunication Engineering from Jadavpur University in 2014.

My research interests broadly span across the domains of multimodal machine learning, computer vision and applications of computer vision in healthcare. I am particularly interested in multimodal learning with major focus along the following dimensions:

  • Robustness: Improving model performance under adversarial settings including modality corruption.
  • Generalization: Enhancing the generalization capabilities of multi-modal models to novel reasoning tasks across diverse groups/cultures and domains.
  • Explainability: Developing interpretable models for multimodal reasoning tasks.

news

Jan 27, 2025 Joined Adobe Research, Bengaluru as a Research Scientist.
Dec 18, 2024 Defended my thesis titled “Multimodal perception guided computational media understanding”
Nov 16, 2023 Our findings on demographic representation in Indian TV shows with Google Research and Geena Davis Institute on Gender in Media published in Google AI India blog and received media coverage
Sep 29, 2023 Passed Ph.D. Qualifying Examination and advanced to candidacy.
Aug 18, 2023 Completed my summer internship at NVIDIA Maxine AI.

selected publications

  1. WACV 2023
    Movieclip: Visual scene recognition in movies
    Digbalay Bose ,  Rajat Hebbar ,  Krishna Somandepalli , and 5 more authors
    In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , 2023
  2. KDD 2023
    FedMultimodal: A Benchmark For Multimodal Federated Learning
    Tiantian Feng ,  Digbalay Bose ,  Tuo Zhang , and 6 more authors
    arXiv preprint arXiv:2306.09486, 2023
  3. FPSAM 2022
    Automatic Analysis of Asymmetry in Facial Paralysis Patients Using Landmark-Based Measures
    Digbalay Bose ,  Krishna Somandepalli ,  Tymon Tai , and 3 more authors
    Facial Plastic Surgery & Aesthetic Medicine, 2022