Department of Computer Science and Engineering · University at Buffalo - SUNY
The Trustworthy and Responsible AI Lab (TRAIL) is part of the CSE department at University at Buffalo-SUNY. Our primary reserach goals are to strive towards building efficient deep learning architectures in domains like privacy-preserving machine learning, generative AI alongside building trustable deep learning systems.
Developing secure machine learning models using Fully Homomorphic Encryption (FHE) to protect data privacy in critical systems like biometrics and healthcare.
Designing hierarchical learning methods and diffusion models to prevent catastrophic forgetting and enhance video and action understanding.
Improving faithfulness in interpretability. Using interpretability techniques to improve robustness, and alignment, diagnose failure modes, and explore its uses in AI4Science.