Thomas Brunner


Hi! I am a researcher with interests in Machine Learning, Computer Graphics and Robotics. Right now, I am finishing my PhD on Adversarial Examples at the chair of Robotics, Artificial Intelligence and Embedded Systems at TU Munich.

I am also a software engineer at X, the moonshot factory (formerly Google X), where I help robots learn stuff!

Find my CV here.

     


News


Publications

These are my first-author publications. For a complete list, see my Google Scholar profile. Please note that these publications were done as part of my PhD work and are not connected to my work at X.



2019


Guessing Smart: Biased Sampling for Efficient Black-Box Adversarial Attacks
Thomas Brunner, Frederik Diehl, Michael Truong Le, Alois Knoll
[Conference paper]

ICCV 2019, Seoul, South Korea
Copy and Paste: A Simple But Effective Initialization Method for Black-Box Adversarial Attacks
Thomas Brunner, Frederik Diehl, Alois Knoll
[Workshop paper]

CVPR 2019, Long Beach, USA
Workshop on Adversarial Machine Learning in Real-World Computer Vision Systems


Leveraging Semantic Embeddings for Safety-Critical Applications
Thomas Brunner, Frederik Diehl, Michael Truong Le, Alois Knoll
[Workshop paper] [Oral presentation]

CVPR 2019, Long Beach, USA
Workshop on Safe Artificial Intelligence for Automated Driving


2018


Guessing Smart: New Directions for Sampling-Based Attacks
Thomas Brunner, Frederik Diehl, Michael Truong Le
[Oral presentation] of our winning submission to the NeurIPS 2018 Adversarial Vision Challenge

NeurIPS 2018, Montréal, Canada
Competition Workshop



Side projects


AlphaSheep: Reinforcement Learning for the Bavarian card game of Schafkopf

My personal playground for learning RL. It's not my main reseach area, but why not have some fun!
Contains a fully-fledged simulator and a graphical user interface. You can play against the agents - try it out!