Sam Pollard (sjpollard)
I am a second year PhD student at the University of Bristol, supervised by Dr. Michael Wray. I’m interested in video redundancy, and exposing the degree to which large vision models are really understanding what they’re evaluating.
Before this, I completed my MEng in Mathematics and Computer Science at the University of Bristol and received a First. My dissertation, titled “Compressed Learning for Egocentric Action Recognition”, was awarded the prize of Best Computer Vision Project.
sam.pollard@bristol.ac.uk, GitHub, Google Scholar
News
[05-06-2025] Video, How Do Your Tokens Merge? now released on arXiv.
[04-06-2025] Video, How Do Your Tokens Merge? code released on GitHub.
[01-04-2025] Video, How Do Your Tokens Merge? accepted at eLVM@CVPR 2025!
[26-02-2025] HD-EPIC: A Highly-Detailed Egocentric Video Dataset accepted at CVPR 2025!
[07-02-2025] HD-EPIC: A Highly-Detailed Egocentric Video Dataset now released on arXiv.
[25-09-2023] Started as a PhD student within the MaVi research group.
[28-07-2023] Graduated from the University of Bristol (MEng) with a First.
Research
Video, How Do Your Tokens Merge?
Sam Pollard, Michael Wray
eLVM@CVPR 2025, Webpage, arXiv
HD-EPIC: A Highly Detailed Egocentric Dataset
Toby Perrett, Ahmad Darkhalil, Saptarshi Sinha, Omar Emara, Sam Pollard, Kranti Parida, Kaiting Liu, Prajwal Gatti, Siddhant Bansal, Kevin Flanagan, Jacob Chalk, Zhifan Zhu, Rhodri Guerrier, Fahd Abdelazim, Bin Zhu, Davide Moltisanti, Michael Wray, Hazel Doughty, Dima Damen
Teaching
Teaching Assistant - COMSM0045 Applied Deep Learning 23/24
Teaching Assistant - COMSM0045 Applied Deep Learning 24/25