Publications

Ireton Liu, Brian Xu, Dominic Stratford, Steven James, Richard Klein, James Tompkin (2026). The Sterkfontein Caves Dataset: A Novel View Synthesis Challenge from the Cradle of Humankind. In European Conference on Computer Vision.

PDF Cite

Tristan Bester, Geraud Nangue Tasse, Benjamin Rosman, Steven James (2026). PyCRM: A Python library for reward machine-based reinforcement learning. SoftwareX.

PDF Cite Code DOI

Steven James (2026). Drowning in Degrees of Freedom: One Agent for Every Task. International Joint Conference on Artificial Intelligence.

PDF Cite

Geraud Nangue Tasse, Steven James, Benjamin Rosman (2026). The Goal-Directed Frame for General Agents. Finding the Frame Workshop at RLC.

PDF Cite

Ebenezer Gelo, Geraud Nangue Tasse, Steven James, Benjamin Rosman (2026). Redistribution-based Cost Inference Improves Sparse Safe Offline RL. Safe Physical AI Workshop at IJCAI.

PDF Cite

Geraud Nangue Tasse, Mark Nemecek, Tamlin Love, Steven James, Benjamin Rosman (2026). An Unreasonably Simple Approach to Safe RL. Reinforcement Learning Journal.

PDF Cite Code

Muhammad Nasir, Yuchen Li, Steven James, Julian Togelius (2026). Mortar: Evolving Mechanics for Automatic Game Design. In Proceedings of the Genetic and Evolutionary Computation Conference.

PDF Cite

Devon Jarvis, Richard Klein, Benjamin Rosman, Steven James, Stefano Sarao Mannelli (2026). The Stochastic Parrot in the Coal Mine: Model Collapse is a Threat to Low-Resource Communities. In International Conference on Machine Learning.

PDF Cite

Damion Harvey, Geraud Nangue Tasse, Benjamin Rosman, Branden Ingram, Steven James (2026). Unsupervised Hierarchical Skill Discovery. In International Conference on Machine Learning.

PDF Cite Code

Caston Nyabadza, Benjamin Rosman, Steven James, Geraud Nangue Tasse (2026). Weighted Composition for Entropy-regularised Reinforcement Learning. ORiON.

PDF Cite

Simon Rosen, Siddarth Singh, Ebenezer Gelo, Helen Sarah Robertson, Ibrahim Suder, Victoria Williams, Benjamin Rosman, Geraud Nangue Tasse, Steven James (2026). MoralityGym: A Benchmark for Evaluating Hierarchical Moral Alignment in Sequential Decision-Making Agents. In International Conference on Autonomous Agents and Multiagent Systems.

PDF Cite

Alper Ahmetoglu, Steven James, Cameron Allen, Sam Lobel, David Abel, George Konidaris (2025). Skill-Driven Neurosymbolic State Abstractions. In Advances in Neural Information Processing Systems.

PDF Cite

Amir Esterhuysen, Steven James, Geraud Nangue Tasse, Benjamin Rosman, Jonathan Shock (2025). Using NEAT to Learn Operators for Flexible Boolean Composition within Reinforcement Learning. The 6th Multi-disciplinary Conference on Reinforcement Learning and Decision Making.

PDF Cite

Ndivhuwo Makondo, Paul Amayo, Jenalea Rajab, Lifhasi Ramulondi, Steven James, Alexandra Barry, Adam Mukuddem, Thabisa Maweni, Tamlin Love, Geraud Nangue Tasse, Zimkhitha Sijovu, Belinda Matebese (2025). Strengthening Robotics and Automation in Africa: Highlights of ICRA@40-Africa. IEEE Robotics & Automation Magazine.

PDF Cite

Geraud Nangue Tasse, Steven James, Benjamin Rosman (2025). Composition and Zero-Shot Transfer with Lattice Structures in Reinforcement Learning. Journal of Artificial Intelligence Research.

PDF Cite DOI

Muhammad Nasir, Steven James, Julian Togelius (2024). GameTraversalBenchmark: Evaluating Planning Abilities Of Large Language Models Through Traversing 2D Game Maps. Advances in Neural Information Processing Systems.

PDF Cite Code

Vanya Cohen, Geraud Nangue Tasse, Nakul Gopalan, Steven James, Matthew Gombolay, Ray Mooney, Benjamin Rosman (2024). Compositional Instruction Following with Language Models and Reinforcement Learning. Transactions on Machine Learning Research.

PDF Cite

Geraud Nangue Tasse, Tamlin Love, Mark Nemecek, Steven James, Benjamin Rosman (2024). ROSARL: Reward-Only Safe Reinforcement Learning. Reinforcement Learning Safety Workshop at RLC.

PDF Cite

Simon Rosen, Abdel Mfougouon Njupoun, Geraud Nangue Tasse, Steven James, Benjamin Rosman (2024). Optimal Task Generalisation in Cooperative Multi-Agent Reinforcement Learning. Coordination and Cooperation in Multi-Agent Reinforcement Learning Workshop at RLC.

PDF Cite

Muhammad Nasir, Sam Earle, Christopher Cleghorn, Steven James, Julian Togelius (2024). LLMatic: Neural Architecture Search via Large Language Models and Quality Diversity Optimization. In Proceedings of the Genetic and Evolutionary Computation Conference.

PDF Cite Code

William Hill, Ireton Liu, Anita De Mello Koch, Damion Harvey, Nishanth Kumar, George Konidaris, Steven James (2024). MinePlanner: A Benchmark for Long-Horizon Planning in Large Minecraft Worlds. 6th ICAPS Workshop on the International Planning Competition.

PDF Cite

Geraud Nangue Tasse, Devon Jarvis, Steven James, Benjamin Rosman (2024). Skill Machines: Temporal Logic Skill Composition in Reinforcement Learning. In Proceedings of the Twelfth International Conference on Learning Representations.

Cite

Tristan Bester, Benjamin Rosman, Steven James, Geraud Nangue Tasse (2024). Counting Reward Automata: Sample Efficient Reinforcement Learning Through the Exploitation of Reward Function Structure. Workshop on Neuro-Symbolic Learning and Reasoning in the Era of Large Language Models at AAAI.

PDF Cite

David Torpey, Max Parkin, Jonah Alter, Richard Klein, Steven James (2024). MiDaS: A Large-Scale Minecraft Dataset for Non-Natural Image Benchmarking. Journal of Electronic Imaging.

PDF Cite DOI

Michael Beukman, Devon Jarvis, Richard Klein, Steven James, Benjamin Rosman (2023). Dynamics Generalisation in Reinforcement Learning via Adaptive Context-Aware Policies. In Advances in Neural Information Processing Systems.

PDF Cite

Eric Rosen, Steven James, Sergio Orozco, Vedant Gupta, Max Merlin, Stefanie Tellex, George Konidaris (2023). Synthesizing Navigation Abstractions for Planning with Portable Manipulation Skills. In Conference on Robot Learning.

PDF Cite

Nathan Michlo, Richard Klein, Steven James (2023). Overlooked Implications of the Reconstruction Loss for VAE Disentanglement. Proceedings of the Thirty-second International Joint Conference on Artificial Intelligence.

PDF Cite Supplementary Material

Michael Beukman, Manuel Fokam, Marcel Kruger, Guy Axelrod, Muhammad Nasir, Branden Ingram, Benjamin Rosman, Steven James (2023). Hierarchically Composing Level Generators for the Creation of Complex Structures. IEEE Transactions on Games.

PDF Cite DOI

Muhammad Nasir, Michael Beukman, Steven James, Christopher Cleghorn (2023). Augmentative Topology Agents For Open-ended Learning. Genetic and Evolutionary Computation Conference Companion.

PDF Cite

Adam Pacheck, Steven James, George Konidaris, Hadas Kress-Gazit (2023). Automatic Encoding and Repair of Reactive High-Level Tasks with Learned Abstract Representations. The International Journal of Robotics Research.

PDF Cite

Vanya Cohen, Geraud Nangue Tasse, Nakul Gopalan, Steven James, Raymond Mooney, Benjamin Rosman (2022). End-to-End Learning to Follow Language Instructions with Compositional Policies. Workshop on Language and Robot Learning @ CORL.

PDF Cite

Geraud Nangue Tasse, Devon Jarvis, Steven James, Benjamin Rosman (2022). Skill Machines: Temporal Logic Composition in Reinforcement Learning. In Deep Reinforcement Learning Workshop @ NeurIPS 2022.

PDF Cite

Geraud Nangue Tasse, Devon Jarvis, Steven James, Benjamin Rosman (2022). Skill Machines: Temporal Logic Composition in Reinforcement Learning. In Lifelong Learning of High-level Cognitive and Reasoning Skills Workshop @ IROS 2022.

PDF Cite Video

Tamlin Love, Devon Jarvis, Geraud Nangue Tasse, Branden Ingram, Steven James, Benjamin Rosman (2022). Facilitating Safe Sim-to-Real through Simulator Abstraction and Zero-shot Task Composition. In Lifelong Learning of High-level Cognitive and Reasoning Skills Workshop @ IROS 2022.

PDF Cite Video

Muhammad Nasir, Steven James, Christopher Cleghorn (2022). Augmentative Topology Agents For Open-ended Learning. In Lifelong Learning of High-level Cognitive and Reasoning Skills Workshop @ IROS 2022.

PDF Cite Video Supplementary Material

Nicholas Muir, Steven James (2022). Combining Evolutionary Search with Behaviour Cloning for Procedurally Generated Content. In Proceedings of 43rd Conference of the South African Institute of Computer Scientists and Information Technologists.

PDF Cite

Michael Beukman, Christopher Cleghorn, Steven James (2022). Procedural Content Generation using Neuroevolution and Novelty Search for Diverse Video Game Levels. In Proceedings of the Genetic and Evolutionary Computation Conference.

PDF Cite Code Video

Geraud Nangue Tasse, Benjamin Rosman, Steven James (2022). World Value Functions: Knowledge Representation for Learning and Planning. Bridging the Gap Between AI Planning and Reinforcement Learning Workshop at ICAPS.

PDF Cite

Geraud Nangue Tasse, Benjamin Rosman, Steven James (2022). World Value Functions: Knowledge Representation for Multitask Reinforcement Learning. The 5th Multi-disciplinary Conference on Reinforcement Learning and Decision Making.

PDF Cite

Steven James, Benjamin Rosman, George Konidaris (2022). Learning Abstract and Transferable Representations for Planning. The 5th Multi-disciplinary Conference on Reinforcement Learning and Decision Making.

PDF Cite

Michael Beukman, Michael Mitcheley, Dean Wookey, Steven James, George Konidaris (2022). Adaptive Online Value Function Approximation with Wavelets. The 5th Multi-disciplinary Conference on Reinforcement Learning and Decision Making.

PDF Cite

Nathan Michlo, Devon Jarvis, Richard Klein, Steven James (2022). Accounting for the Sequential Nature of States to Learn Representations in Reinforcement Learning. The 5th Multi-disciplinary Conference on Reinforcement Learning and Decision Making.

PDF Cite

Geraud Nangue Tasse, Steven James, Benjamin Rosman (2022). Generalisation in Lifelong Reinforcement Learning through Logical Composition. In Proceedings of the Tenth International Conference on Learning Representations.

PDF Cite

Steven James, Benjamin Rosman, George Konidaris (2022). Autonomous Learning of Object-Centric Abstractions for High-Level Planning. In Proceedings of the Tenth International Conference on Learning Representations.

PDF Cite

Nishai Kooverjee, Steven James, Terence Van Zyl (2022). Investigating Transfer Learning in Graph Neural Networks. Electronics.

PDF Cite

Geraud Nangue Tasse, Steven James, Benjamin Rosman (2021). Generalisation in Lifelong Reinforcement Learning through Logical Composition. NeurIPS Deep Reinforcement Learning Workshop.

PDF Cite

Vanya Cohen, Geraud Nangue Tasse, Nakul Gopalan, Steven James, Matthew Gombolay, Benjamin Rosman (2021). Learning to Follow Language Instructions with Compositional Policies. AAAI Fall Symposium on AI for Human-Robot Interaction.

PDF Cite

Geraud Nangue Tasse, Steven James, Benjamin Rosman (2020). A Boolean Task Algebra for Reinforcement Learning. In Advances in Neural Information Processing Systems.

PDF Cite

Geraud Nangue Tasse, Steven James, Benjamin Rosman (2020). Logical Composition for Lifelong Reinforcement Learning. 4th Lifelong Learning Workshop at ICML.

PDF Cite

Steven James, Benjamin Rosman, George Konidaris (2020). Learning Portable Representations for High-Level Planning. International Conference on Machine Learning.

PDF Cite

Steven James, Benjamin Rosman, George Konidaris (2020). Learning Object-Centric Representations for High-Level Planning in Minecraft. Object-Oriented Learning: Perception, Representation, and Reasoning. Workshop at ICML.

PDF Cite

Arnu Pretorius, Elan Van Biljon, Benjamin Van Niekerk, Ryan Eloff, Matthew Reynard, Steven James, Benjamin Rosman, Herman Kamper, Steve Kroon (2020). If Dropout Limits Trainable Depth, Does Critical Initialisation Still Matter? A Large-scale Statistical Analysis on ReLU Networks. Pattern Recognition Letters.

PDF Cite

Geraud Nangue Tasse, Steven James, Benjamin Rosman (2020). A Boolean Task Algebra for Reinforcement Learning. Beyond Tabula Rasa in Reinforcement Learning (Workshop at ICLR).

PDF Cite

Kimessha Paupamah, Steven James, Richard Klein (2020). Quantisation and Pruning for Neural Network Compression and Regularisation. International SAUPEC/RobMech/PRASA Conference.

PDF Cite

Matthew Cockcroft, Shahil Mawjee, Steven James, Pravesh Ranchod (2020). Learning Options from Demonstration using Skill Segmentation. International SAUPEC/RobMech/PRASA Conference.

PDF Cite

Nishai Kooverjee, Steven James, Terence Van Zyl (2020). Inter-and Intra-domain Knowledge Transfer for Related Tasks in Deep Character Recognition. International SAUPEC/RobMech/PRASA Conference.

PDF Cite

Benjamin Van Niekerk, Steven James, Adam Earle, Benjamin Rosman (2019). Composing Value Functions in Reinforcement Learning. International Conference on Machine Learning.

PDF Cite Supplementary Material

Craig Bester, Steven James, George Konidaris (2019). Multi-Pass Q-Networks for Deep Reinforcement Learning with Parameterised Action Spaces. Technical Report.

PDF Cite

Benjamin Van Niekerk, Steven James, Adam Earle, Benjamin Rosman (2018). Will it Blend? Composing Value Functions in Reinforcement Learning. The 2nd Lifelong Learning: A Reinforcement Learning Approach (LLARLA) Workshop @ FAIM.

PDF Cite

Steven James, Benjamin Rosman, George Konidaris (2018). Learning to Plan with Portable Symbols. ICML/IJCAI/AAMAS 2018 Workshop on Planning and Learning.

PDF Cite

Steven James, George Konidaris, Benjamin Rosman (2017). An Analysis of Monte Carlo Tree Search. Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence.

PDF Cite

Steven James, Benjamin Rosman, George Konidaris (2016). An Investigation into the Effectiveness of Heavy Rollouts in UCT. General Intelligence in Game-Playing Agents Workshop at IJCAI.

PDF Cite