I am a computer scientist and mathematician from Salt Lake City, Utah. I’m currently doing my PhD in deep learning theory and deep reinforcement learning at Carnegie Mellon University under Ruslan Salakhutdinov. A large portion of my time is spent working on MineRL, a project I co-founded at CMU, with the broad goal of developing general AI in Minecraft using human priors.
My goal is to understand human intelligence by creating artificial intelligence.
On Universal Approximation by Neural Networks with Uniform Guarantees on Approximation of Infinite Dimensional Maps William H. Guss, Ruslan Salakhutdinov. Preprint. [arXiv]
MineRL: A large-scale dataset of Minecraft demonstrations. William H. Guss, Brandon Houghton, Nicholay Topin, Phillip Wang, Cayden Codel, Manuela Veloso, Ruslan Salakhutdinov. IJCAI, 2019. [arXiv]
The MineRL competition on sample efficient reinforcement learning using human priors. William H. Guss, Cayden Codel, Katja Hofmann, Brandon Houghton, Noboru Kuno, Stephanie Milani, Sharada Mohanty, Diego Perez Liebana, Ruslan Salakhutdinov, Nicholay Topin, Manuela Veloso, Phillip Wang. NeurIPS Competition Track, 2019. [arXiv]
Towards a theory of neural topology: The homology of rectified units. William H. Guss and Ruslan Salakhutdinov. ICML 2018 GIML Workshlop. Best Contribution Award. [extended abstract]
On Characterizing the Capacity of Neural Networks using Algebraic Topology. William H. Guss, Ruslan Salakhutdinov. NIPS 2017, DLTP Workshop. [arXiv] [poster]
Towards Neural Homology Theory. William H. Guss, Ruslan Salakhutdinov. Talk, Microsoft Research, 2018. [slides]
Eigen: A Step Towards Conversational AI. William H. Guss, James Bartlett, Phillip Kuznetsov, Piyush Patil. Alexa Prize Proceedings 2017. [proceedings]
Deep Function Machines: Generalized Neural Networks for Topological Layer Expression. William H. Guss. Preprint. [arXiv]
Universal Approximation of Nonlinear Operators on Banach Space. William H. Guss. Machine Learning at Berkeley Research Symposium 2016. [pdf]
Backpropagation-Free Parallel Deep Reinforcement Learning. William H. Guss. James Bartlett, Noah Golmant, Phillip Kuznetsov, Max Johansen. Preprint (WIP). [pdf]
Parameter Reduction using Operator Neural Networks. William H. Guss. Microsoft Research Symposium 2016. Best Poster Award. [poster]
Functional Neural Networks Evaluated by Weierstrass Polynomials. William H. Guss, Phillip Kuznetsov, Patrick Chen. Intel ISEF 2015. Pittsburgh, Pennsylvania. [AAAI’ Honorable Mention] [ASA’ Honorable Mention]