Publications | Steve Jameshttps://sdjames.me/publication/PublicationsHugo Blox Builder (https://hugoblox.com)en-ushttps://sdjames.me/media/icon_huf407e28faea15ab880cdbc1ef97b5c3b_31359_512x512_fill_lanczos_center_3.pngPublicationshttps://sdjames.me/publication/The Sterkfontein Caves Dataset: A Novel View Synthesis Challenge from the Cradle of Humankindhttps://sdjames.me/publication/liu-2026-sterkfontein/Wed, 09 Sep 2026 00:00:00 +0000https://sdjames.me/publication/liu-2026-sterkfontein/PyCRM: A Python library for reward machine-based reinforcement learninghttps://sdjames.me/publication/bester-2026-pycrm/Tue, 01 Sep 2026 00:00:00 +0000https://sdjames.me/publication/bester-2026-pycrm/Drowning in Degrees of Freedom: One Agent for Every Taskhttps://sdjames.me/publication/james-2026-drowning/Wed, 19 Aug 2026 00:00:00 +0000https://sdjames.me/publication/james-2026-drowning/Redistribution-based Cost Inference Improves Sparse Safe Offline RLhttps://sdjames.me/publication/gelo-2026-redistribution/Sun, 16 Aug 2026 00:00:00 +0000https://sdjames.me/publication/gelo-2026-redistribution/The Goal-Directed Frame for General Agentshttps://sdjames.me/publication/nangue-2026-goal/Sun, 16 Aug 2026 00:00:00 +0000https://sdjames.me/publication/nangue-2026-goal/An Unreasonably Simple Approach to Safe RLhttps://sdjames.me/publication/nangue-2026-unreasonably/Sat, 15 Aug 2026 00:00:00 +0000https://sdjames.me/publication/nangue-2026-unreasonably/Mortar: Evolving Mechanics for Automatic Game Designhttps://sdjames.me/publication/nasir-2026-mortar/Mon, 13 Jul 2026 00:00:00 +0000https://sdjames.me/publication/nasir-2026-mortar/The Stochastic Parrot in the Coal Mine: Model Collapse is a Threat to Low-Resource Communitieshttps://sdjames.me/publication/jarvis-2026-position/Tue, 07 Jul 2026 00:00:00 +0000https://sdjames.me/publication/jarvis-2026-position/Unsupervised Hierarchical Skill Discoveryhttps://sdjames.me/publication/harvey-2026-unsupervised/Mon, 06 Jul 2026 00:00:00 +0000https://sdjames.me/publication/harvey-2026-unsupervised/Weighted Composition for Entropy-regularised Reinforcement Learninghttps://sdjames.me/publication/nyabadza-2026-weighted/Wed, 01 Jul 2026 00:00:00 +0000https://sdjames.me/publication/nyabadza-2026-weighted/MoralityGym: A Benchmark for Evaluating Hierarchical Moral Alignment in Sequential Decision-Making Agentshttps://sdjames.me/publication/rosen-2026-moralitygym/Mon, 25 May 2026 00:00:00 +0000https://sdjames.me/publication/rosen-2026-moralitygym/Skill-Driven Neurosymbolic State Abstractionshttps://sdjames.me/publication/ahmetoglu-2025-skill/Thu, 04 Dec 2025 00:00:00 +0000https://sdjames.me/publication/ahmetoglu-2025-skill/Using NEAT to Learn Operators for Flexible Boolean Composition within Reinforcement Learninghttps://sdjames.me/publication/esterhuysen-2025-using/Wed, 11 Jun 2025 00:00:00 +0000https://sdjames.me/publication/esterhuysen-2025-using/Strengthening Robotics and Automation in Africa: Highlights of ICRA@40-Africahttps://sdjames.me/publication/makondo-2025-strengthening/Sun, 01 Jun 2025 00:00:00 +0000https://sdjames.me/publication/makondo-2025-strengthening/Composition and Zero-Shot Transfer with Lattice Structures in Reinforcement Learninghttps://sdjames.me/publication/nangue-2025-composition/Thu, 17 Apr 2025 00:00:00 +0000https://sdjames.me/publication/nangue-2025-composition/GameTraversalBenchmark: Evaluating Planning Abilities Of Large Language Models Through Traversing 2D Game Mapshttps://sdjames.me/publication/nasir-2024-gametraversalbenchmark/Tue, 10 Dec 2024 00:00:00 +0000https://sdjames.me/publication/nasir-2024-gametraversalbenchmark/Compositional Instruction Following with Language Models and Reinforcement Learninghttps://sdjames.me/publication/cohen-2024-compositional/Sun, 01 Dec 2024 00:00:00 +0000https://sdjames.me/publication/cohen-2024-compositional/Optimal Task Generalisation in Cooperative Multi-Agent Reinforcement Learninghttps://sdjames.me/publication/rosen-2024-optimal/Fri, 09 Aug 2024 00:00:00 +0000https://sdjames.me/publication/rosen-2024-optimal/ROSARL: Reward-Only Safe Reinforcement Learninghttps://sdjames.me/publication/nangue-2024-rosarl/Fri, 09 Aug 2024 00:00:00 +0000https://sdjames.me/publication/nangue-2024-rosarl/LLMatic: Neural Architecture Search via Large Language Models and Quality Diversity Optimizationhttps://sdjames.me/publication/nasir-2024-llmatic/Sun, 14 Jul 2024 00:00:00 +0000https://sdjames.me/publication/nasir-2024-llmatic/MinePlanner: A Benchmark for Long-Horizon Planning in Large Minecraft Worldshttps://sdjames.me/publication/hill-2024-mineplanner/Sun, 02 Jun 2024 00:00:00 +0000https://sdjames.me/publication/hill-2024-mineplanner/Skill Machines: Temporal Logic Skill Composition in Reinforcement Learninghttps://sdjames.me/publication/nangue-2024-skill/Tue, 07 May 2024 00:00:00 +0000https://sdjames.me/publication/nangue-2024-skill/Counting Reward Automata: Sample Efficient Reinforcement Learning Through the Exploitation of Reward Function Structurehttps://sdjames.me/publication/bester-2024-counting/Mon, 26 Feb 2024 00:00:00 +0000https://sdjames.me/publication/bester-2024-counting/MiDaS: A Large-Scale Minecraft Dataset for Non-Natural Image Benchmarkinghttps://sdjames.me/publication/torpey-2024-midas/Mon, 05 Feb 2024 00:00:00 +0000https://sdjames.me/publication/torpey-2024-midas/Dynamics Generalisation in Reinforcement Learning via Adaptive Context-Aware Policieshttps://sdjames.me/publication/beukman-2023-dynamics/Mon, 04 Dec 2023 00:00:00 +0000https://sdjames.me/publication/beukman-2023-dynamics/Synthesizing Navigation Abstractions for Planning with Portable Manipulation Skillshttps://sdjames.me/publication/rosen-2023-synthesizing/Mon, 06 Nov 2023 00:00:00 +0000https://sdjames.me/publication/rosen-2023-synthesizing/Overlooked Implications of the Reconstruction Loss for VAE Disentanglementhttps://sdjames.me/publication/michlo-2023-overlooked/Sat, 26 Aug 2023 00:00:00 +0000https://sdjames.me/publication/michlo-2023-overlooked/Hierarchically Composing Level Generators for the Creation of Complex Structureshttps://sdjames.me/publication/beukman-2023-hierarchically/Fri, 21 Jul 2023 00:00:00 +0000https://sdjames.me/publication/beukman-2023-hierarchically/Augmentative Topology Agents For Open-ended Learninghttps://sdjames.me/publication/nasir-2023-augmentative/Sat, 15 Jul 2023 00:00:00 +0000https://sdjames.me/publication/nasir-2023-augmentative/Automatic Encoding and Repair of Reactive High-Level Tasks with Learned Abstract Representationshttps://sdjames.me/publication/pacheck-2023-automatic/Tue, 23 May 2023 00:00:00 +0000https://sdjames.me/publication/pacheck-2023-automatic/End-to-End Learning to Follow Language Instructions with Compositional Policieshttps://sdjames.me/publication/cohen-2022-end/Wed, 14 Dec 2022 00:00:00 +0000https://sdjames.me/publication/cohen-2022-end/Skill Machines: Temporal Logic Composition in Reinforcement Learninghttps://sdjames.me/publication/nangue-2022-skill-b/Fri, 09 Dec 2022 00:00:00 +0000https://sdjames.me/publication/nangue-2022-skill-b/Augmentative Topology Agents For Open-ended Learninghttps://sdjames.me/publication/nasir-2022-augmentative/Sun, 23 Oct 2022 00:00:00 +0000https://sdjames.me/publication/nasir-2022-augmentative/Facilitating Safe Sim-to-Real through Simulator Abstraction and Zero-shot Task Compositionhttps://sdjames.me/publication/love-2022-facilitating/Sun, 23 Oct 2022 00:00:00 +0000https://sdjames.me/publication/love-2022-facilitating/Skill Machines: Temporal Logic Composition in Reinforcement Learninghttps://sdjames.me/publication/nangue-2022-skill-a/Sun, 23 Oct 2022 00:00:00 +0000https://sdjames.me/publication/nangue-2022-skill-a/Combining Evolutionary Search with Behaviour Cloning for Procedurally Generated Contenthttps://sdjames.me/publication/muir-2022-combining/Mon, 18 Jul 2022 00:00:00 +0000https://sdjames.me/publication/muir-2022-combining/Procedural Content Generation using Neuroevolution and Novelty Search for Diverse Video Game Levelshttps://sdjames.me/publication/beukman-2022-procedural/Sat, 09 Jul 2022 00:00:00 +0000https://sdjames.me/publication/beukman-2022-procedural/World Value Functions: Knowledge Representation for Learning and Planninghttps://sdjames.me/publication/nangue-2022-world-b/Mon, 13 Jun 2022 00:00:00 +0000https://sdjames.me/publication/nangue-2022-world-b/Accounting for the Sequential Nature of States to Learn Representations in Reinforcement Learninghttps://sdjames.me/publication/michlo-2022-accounting/Wed, 08 Jun 2022 00:00:00 +0000https://sdjames.me/publication/michlo-2022-accounting/Adaptive Online Value Function Approximation with Waveletshttps://sdjames.me/publication/beukman-2022-adaptive/Wed, 08 Jun 2022 00:00:00 +0000https://sdjames.me/publication/beukman-2022-adaptive/Learning Abstract and Transferable Representations for Planninghttps://sdjames.me/publication/james-2022-learning/Wed, 08 Jun 2022 00:00:00 +0000https://sdjames.me/publication/james-2022-learning/World Value Functions: Knowledge Representation for Multitask Reinforcement Learninghttps://sdjames.me/publication/nangue-2022-world-a/Wed, 08 Jun 2022 00:00:00 +0000https://sdjames.me/publication/nangue-2022-world-a/Autonomous Learning of Object-Centric Abstractions for High-Level Planninghttps://sdjames.me/publication/james-2022-autonomous/Mon, 25 Apr 2022 00:00:00 +0000https://sdjames.me/publication/james-2022-autonomous/Generalisation in Lifelong Reinforcement Learning through Logical Compositionhttps://sdjames.me/publication/nangue-2022-generalisation/Mon, 25 Apr 2022 00:00:00 +0000https://sdjames.me/publication/nangue-2022-generalisation/Investigating Transfer Learning in Graph Neural Networkshttps://sdjames.me/publication/kooverjee-2022-investigating/Sat, 09 Apr 2022 00:00:00 +0000https://sdjames.me/publication/kooverjee-2022-investigating/Generalisation in Lifelong Reinforcement Learning through Logical Compositionhttps://sdjames.me/publication/nangue-2021-generalisation/Mon, 06 Dec 2021 00:00:00 +0000https://sdjames.me/publication/nangue-2021-generalisation/Learning to Follow Language Instructions with Compositional Policieshttps://sdjames.me/publication/cohen-2021-learning/Thu, 04 Nov 2021 00:00:00 +0000https://sdjames.me/publication/cohen-2021-learning/A Boolean Task Algebra for Reinforcement Learninghttps://sdjames.me/publication/nangue-2020-boolean-a/Sun, 06 Dec 2020 00:00:00 +0000https://sdjames.me/publication/nangue-2020-boolean-a/Logical Composition for Lifelong Reinforcement Learninghttps://sdjames.me/publication/nangue-2020-logical/Mon, 13 Jul 2020 00:00:00 +0000https://sdjames.me/publication/nangue-2020-logical/Learning Portable Representations for High-Level Planninghttps://sdjames.me/publication/james-2020-learning-a/Sun, 12 Jul 2020 00:00:00 +0000https://sdjames.me/publication/james-2020-learning-a/Learning Object-Centric Representations for High-Level Planning in Minecrafthttps://sdjames.me/publication/james-2020-learning-b/Sat, 11 Jul 2020 00:00:00 +0000https://sdjames.me/publication/james-2020-learning-b/If Dropout Limits Trainable Depth, Does Critical Initialisation Still Matter? A Large-scale Statistical Analysis on ReLU Networkshttps://sdjames.me/publication/pretorius-2020-if/Sun, 28 Jun 2020 00:00:00 +0000https://sdjames.me/publication/pretorius-2020-if/A Boolean Task Algebra for Reinforcement Learninghttps://sdjames.me/publication/nangue-2020-boolean-b/Sun, 26 Apr 2020 00:00:00 +0000https://sdjames.me/publication/nangue-2020-boolean-b/Inter-and Intra-domain Knowledge Transfer for Related Tasks in Deep Character Recognitionhttps://sdjames.me/publication/kooverjee-2020-inter/Wed, 29 Jan 2020 00:00:00 +0000https://sdjames.me/publication/kooverjee-2020-inter/Learning Options from Demonstration using Skill Segmentationhttps://sdjames.me/publication/cockcroft-2020-learning/Wed, 29 Jan 2020 00:00:00 +0000https://sdjames.me/publication/cockcroft-2020-learning/Quantisation and Pruning for Neural Network Compression and Regularisationhttps://sdjames.me/publication/paupamah-2020-quantisation/Wed, 29 Jan 2020 00:00:00 +0000https://sdjames.me/publication/paupamah-2020-quantisation/Composing Value Functions in Reinforcement Learninghttps://sdjames.me/publication/van-2019-composing/Mon, 10 Jun 2019 00:00:00 +0000https://sdjames.me/publication/van-2019-composing/Multi-Pass Q-Networks for Deep Reinforcement Learning with Parameterised Action Spaceshttps://sdjames.me/publication/bester-2019-multi/Fri, 10 May 2019 00:00:00 +0000https://sdjames.me/publication/bester-2019-multi/Learning to Plan with Portable Symbolshttps://sdjames.me/publication/james-2018-learning/Sat, 14 Jul 2018 00:00:00 +0000https://sdjames.me/publication/james-2018-learning/Will it Blend? Composing Value Functions in Reinforcement Learninghttps://sdjames.me/publication/van-2018-will/Sat, 14 Jul 2018 00:00:00 +0000https://sdjames.me/publication/van-2018-will/An Analysis of Monte Carlo Tree Searchhttps://sdjames.me/publication/james-2017-analysis/Sat, 04 Feb 2017 00:00:00 +0000https://sdjames.me/publication/james-2017-analysis/An Investigation into the Effectiveness of Heavy Rollouts in UCThttps://sdjames.me/publication/james-2016-investigation/Sat, 09 Jul 2016 00:00:00 +0000https://sdjames.me/publication/james-2016-investigation/