Source Models Leak What They Shouldn't: Unlearning Zero-shot Transfer in Domain Adaptation Through Adversarial Optimization
CVPR 2026
Poornima S Jain
I am a Ph.D. student in the Department of Artificial Intelligence at IIT Hyderabad under the supervision of Dr. Vineeth Balasubramanian. My Ph.D. research area is Privacy and Safety in AI with a focus on Machine Unlearning.
Previously, I completed my MS-by-Research on Complex-valued Neural Networks for MRI Reconstruction at IIIT-Bangalore with Dr. Neelam Sinha and Dr. G. Srinivasaraghavan, and B.Tech in Computer Science Engineering from VNIT-Nagpur.
Research Interests: AI Safety, Privacy and Alignment; Multimodal LLMs; Machine Unlearning; Domain Adaptation
Latest News
- MAR 2026 Paper accepted at the CVPR 2026 Workshop on New Trends in AI-Generated Media and Security!
- FEB 2026 Paper accepted at CVPR 2026 (Main track), Denver Colorado!
- DEC 2025 Presented poster at IndoML-2025 held at BITS-Pilani Hyderabad.
- DEC 2024 Selected to participate in IndoML-2024.
- JUL 2024 Joined Ph.D. at the Machine Learning and Vision Lab, IIT Hyderabad.
- JUL 2023 Won first place at the Student Research Symposium at NCVPRIPG-2023, IIT Jodhpur!
- JUL 2023 Paper accepted in short paper track at MIDL 2023, Vanderbilt University, Nashville!
- DEC 2022 Paper presented at ICVGIP 2022, IIT Gandhinagar
- JUL 2022 Paper accepted at EMBC 2022, Glasgow, UK!
Publications
An end-to-end Complex-valued Neural Network approach for k-space interpolation in Parallel MRI
MIDL 2023
Volunteering & Teaching
-
DEC 2023
Volunteer
IEEE International Conference on High Performance Computing, Data, and Analytics (HiPC), Bangalore, India. -
Jan - May 2022
Teaching Assistant
SM202-Maths 4: Probability, Statistics and Complex Analysis for 2nd year Integrated MTech.