Pranav Kulkarni

Incoming CS Ph.D. Student @ UMD

prof_pic.jpg

Hi! I am an incoming Ph.D. student at the University of Maryland, advised by Heng Huang. Currently, I work as a software engineer at the University of Maryland Institute for Health Computing (UM-IHC). Previously, I was a full-time researcher at the University of Maryland Medical Intelligent Imaging (UM2ii) Center, where I was advised by Vishwa Parekh and Paul Yi. Prior to that, I received dual B.S. degrees in Computer Science and Mathematics from the University of Maryland in 2022.

My research is primarly focused on the intersection of machine learning, computer vision, and medical imaging, with the goal of improving healthcare outcomes. My current research interests include:

  1. Multi-modal models that integrate imaging, clinical, and multi-omics data to enable opportunistic screening for early-stage disease detection.

  2. Federated learning techniques that leverage distributed, heterogeneous data to reduce burden of medical image annotation in a privacy-preserving way.

  3. Trustworthy and explainable AI to adapt to distribution shifts over time and mitigate algorithmic bias.

Furthermore, I have ongoing collaborations with clinicians in translational research to bring cutting-edge AI research from bench to bedside. In my free time, I enjoy reading about history, collecting vintage maps, gardening, and hiking.

Selected Publications

  1. From Isolation to Collaboration: Federated Class-Heterogeneous Learning for Chest X-Ray Classification
    Pranav Kulkarni, Adway Kanhere, Paul H. Yi, and Vishwa S. Parekh
    In AHLI Machine Learning for Health Symposium (ML4H), 623–635. Dec 2024
    HTML PDF Code arXiv
  2. Spotlight Talk
    Hidden in Plain Sight: Undetectable Adversarial Bias Attacks on Vulnerable Patient Populations
    Pranav Kulkarni, Andrew Chan, Nithya Navarathna, Skylar Chan, Paul H. Yi, and Vishwa S. Parekh
    In Medical Imaging with Deep Learning (MIDL), 793–821. Jul 2024
    HTML PDF Code arXiv
  3. ISLE: An Intelligent Streaming Framework for High-Throughput AI Inference in Medical Imaging
    Pranav Kulkarni, Adway Kanhere, Eliot L. Siegel, Paul H. Yi, and Vishwa S. Parekh
    Journal of Imaging Informatics in Medicine, 37(6), 3250–3263. Jun 2024
    HTML Code arXiv
  4. Privacy-Preserving Collaboration for Multi-Organ Segmentation via Federated Learning from Sites with Partial Labels
    Adway Kanhere, Pranav Kulkarni, Paul H. Yi, and Vishwa S. Parekh
    In CVPR Workshop on Data Curation and Augmentation in Medical Imaging (CVPR DCA-in-MI), 2380–2387. Jun 2024
    HTML PDF Code arXiv
  5. Economic and environmental costs of cloud for medical imaging and radiology artificial intelligence
    Florence X. Doo, Pranav Kulkarni, Eliot Siegel, Michael Toland, Paul H. Yi, Ruth C. Carlos, and Vishwa S. Parekh
    Journal of the American College of Radiology, 21(2), 248–256. Feb 2024
    HTML PDF