Steve(n) James

Senior Lecturer, University of the Witwatersrand
I'm a senior lecturer and member of the RAIL Lab research group at the University of the Witwatersrand in Johannesburg, South Africa. My research interests revolve around artificial intelligence, and reinforcement learning in particular. I'm currently interested in developing agents capable of learning symbolic state representations that can be transferred between tasks. I count myself extremely fortunate to be have been supervised by George Konidaris and Benjamin Rosman.

Latest News

Our paper, Procedural Content Generation using Neuroevolution and Novelty Search for Diverse Video Game Levels, was accepted to the 2022 Genetic and Evolutionary Computation Conference.

We had two papers, Generalisation in Lifelong Reinforcement Learning through Logical Composition and Autonomous Learning of Object-Centric Abstractions for High-Level Planning, accepted to the 2022 International Conference on Learning Representations.

February 2022

Our paper, A Boolean Task Algebra for Reinforcement Learning, was accepted to the 2020 Conference on Neural Information Processing Systems (NeurIPS).

Our paper, Learning Portable Representations for High-Level Planning, was accepted to the 2020 International Conference on Machine Learning.

Our paper, Composing Value Functions in Reinforcement Learning, was accepted to the 2019 International Conference on Machine Learning.

Latest Blog Posts

Composing Value Functions in Reinforcement Learning

  • 11 May 2019
In this work, we investigate ways an agent can combine existing skills to create novel ones in a manner that is both principled and optimal. We find that by constraining the reward function and tra...