Low-Rank Adaptation
Hu et al., ICLR 2022
Teach a frozen model a new task by learning a small, low-rank change to its weights, and how the idea carries over to vision transformers.
Paper slides
Slides for the papers I present in our reading group at the University of Rochester.
October 2026
Visual Prompt Tuning
Jia et al., ECCV 2022
Keep a pretrained vision transformer frozen and learn a few extra input tokens for each new task. With my own runs on three medical image sets and a live demo.
VPT and LoRA
Jia et al., ECCV 2022 · Hu et al., ICLR 2022
Two ways to adapt a frozen vision transformer to a new task, learned input vectors and a small low-rank change to the weights, with my own runs of both on three medical image sets and a live demo.
September 2026
Masked Autoencoders Are Scalable Vision Learners
He et al., CVPR 2022
Hide three quarters of an image and train a vision transformer to paint the missing patches back. What the encoder learns along the way transfers well to other tasks.