Background

Hey

My research interests consist of learning how to move by utilizing simulation.
Trying to develop simple ideas that work well.

Currently computer science PhD student at UT Austin advised by Philipp Krähenbühl
Previously an Intel AI Resident advised by Vladlen Koltun.

Email: brady.zhou (at) utexas.edu
Desk: GDC 4.702E

   

Research

2019
Learning by Cheating
Dian Chen, Brady Zhou, Vladlen Koltun, Philipp Krähenbühl
CoRL 2019, Spotlight
[pdf] [code]

Learning from an agent trained on ground truth vision provides strong supervision and performs well in CARLA.
Does computer vision matter for action?
Brady Zhou, Philipp Krähenbühl, Vladlen Koltun
Science Robotics 2019, Special Section on Computer Vision
[link] [pdf] [code]

Controlled experiments indicate that explicit intermediate representations help action.
Don't let your Discriminator be fooled
Brady Zhou, Philipp Krähenbühl
ICLR 2019
[link] [pdf] [poster] [code]

A discriminator that is not easily fooled by adversarial examples makes GAN training more robust and leads to a smoother objective.
2018
GPU accelerated k-nearest neighbor kernel for sparse feature datasets
Brady Zhou, George Biros
UT MS Research
[link] [pdf] [code]

KNN on sparse features can be accelerated on the GPU using sparse GEMM routines and parallelized bitonic sort.

Teaching

Fall 2020 - Assistant Instructor for CS342: Neural Networks
Summer 2020 - Teaching Assistant for CS395T: Deep Learning
Fall 2019 - Teaching Assistant for CS342: Neural Networks
Fall 2016 - Undergraduate Teaching Assistant for CS312: Introduction to Programming

Other

When I’m not working, I’m probably watching Twitch.tv or playing video games.

I also have two cat roommates - Pig and Teddy.