Ali Hassani

CS PhD student at University of Oregon

I am a computer science PhD student at University of Oregon, and am advised by Prof. Humphrey Shi. My research interests are attention-based models, representation learning, and self-supervision. During my undergrad, I did research under the supervision of Prof. Abbas Salemi at the Mahani Mathematical Research Center.

CV / GitHub / Google Scholar / Twitter

News

  • We recently released Neighborhood Attention Transformer (NAT), a hierarchical attention-based model for vision, built from a localized neighbor-based attention mechanism, along with its customized CUDA kernel.

Select publications & preprints

2022

2021

2021

Research Experience

Since 02/2021

SHI Lab @ University of Oregon.

I've been working on attention-based models for vision (i.e. vision transformers) and investigating data-efficiency. We recently released NAT, a model based on Neighborhood Attention, which competes with state-of-the-art attention-based models (Swin) and CNNs (ConvNeXt). Our first work, CCT extended vision transformers to data-limited regimes, and competed with state-of-the-art CNNs on datasets such as CIFAR-10. It also became the new state of the art on Flowers-102.

2019 to 2021

Mahani Mathematical Research Center.

My research consisted of clustering and feature extraction using matrix factorization.

Teaching

Winter 2022

Spring 2021