Arnu Pretorius — accepted AI-conference papers

2025

Breaking the Performance Ceiling in Reinforcement Learning requires Inference Strategies

Felix Chalumeau, Daniel Rajaonarivonivelomanantsoa, Ruan John de Kock, Juan Claude Formanek, Sasha Abramowitz, Omayma Mahjoub +8

NeurIPS 2025oral

Reinforcement learning (RL) systems have countless applications, from energy-grid management to protein design. However, such real-world scenarios are often extremely difficult, combinatorial in nature, and require complex coordination between multiple agents. This level of complexity can cause even…

Cited by 0 Source ↗Scholar ↗

2025

Memory-Enhanced Neural Solvers for Routing Problems

Felix Chalumeau, Refiloe Shabe, Noah De Nicola, Arnu Pretorius, Thomas D Barrett, Nathan Grinsztajn

NeurIPS 2025spotlight

Routing Problems are central to many real-world applications, yet remain challenging due to their (NP-)hard nature. Amongst existing approaches, heuristics often offer the best trade-off between quality and scalability, making them suitable for industrial use. While Reinforcement Learning (RL) offer…

Cited by 0 Source ↗Scholar ↗

2025

Oryx: a Scalable Sequence Model for Many-Agent Coordination in Offline MARL

Juan Claude Formanek, Omayma Mahjoub, Louay Ben Nessir, Sasha Abramowitz, Ruan John de Kock, Wiem Khlifi +7

NeurIPS 2025poster

A key challenge in offline multi-agent reinforcement learning (MARL) is achieving effective many-agent multi-step coordination in complex environments. In this work, we propose Oryx, a novel algorithm for offline cooperative MARL to directly address this challenge. Oryx adapts the recently proposed…

Cited by 0 Source ↗Scholar ↗

2025

Sable: a Performant, Efficient and Scalable Sequence Model for MARL

Omayma Mahjoub, Sasha Abramowitz, Ruan John de Kock, Wiem Khlifi, Simon Verster Du Toit, Jemma Daniel +5

ICML 2025poster

As multi-agent reinforcement learning (MARL) progresses towards solving larger and more complex problems, it becomes increasingly important that algorithms exhibit the key properties of (1) strong performance, (2) memory efficiency, and (3) scalability. In this work, we introduce Sable, a performant…

Cited by 9 Source ↗PDF ↗code ↗Scholar ↗

2024

Dispelling the Mirage of Progress in Offline MARL through Standardised Baselines and Evaluation

Juan Claude Formanek, Callum Rhys Tilbury, Louise Beyers, Jonathan Phillip Shock, Arnu Pretorius

NeurIPS 2024poster

Offline multi-agent reinforcement learning (MARL) is an emerging field with great promise for real-world applications. Unfortunately, the current state of research in offline MARL is plagued by inconsistencies in baselines and evaluation protocols, which ultimately makes it difficult to accurately a…

Cited by 1 Source ↗PDF ↗code ↗Scholar ↗

2024

Jumanji: a Diverse Suite of Scalable Reinforcement Learning Environments in JAX

Clément Bonnet, Daniel Luo, Donal John Byrne, Shikha Surana, Sasha Abramowitz, Paul Duckworth +18

ICLR 2024poster

Open-source reinforcement learning (RL) environments have played a crucial role in driving progress in the development of AI algorithms. In modern RL research, there is a need for simulated environments that are performant, scalable, and modular to enable their utilization in a wider range of potent…

Cited by 27 Source ↗PDF ↗code ↗Scholar ↗

2024

Should we be going MAD? A Look at Multi-Agent Debate Strategies for LLMs

Andries Petrus Smit, Nathan Grinsztajn, Paul Duckworth, Thomas D Barrett, Arnu Pretorius

ICML 2024poster

Recent advancements in large language models (LLMs) underscore their potential for responding to inquiries in various domains. However, ensuring that generative agents provide accurate and reliable answers remains an ongoing challenge. In this context, multi-agent debate (MAD) has emerged as a promi…

Cited by 15 Source ↗PDF ↗code ↗Scholar ↗

2023

Combinatorial Optimization with Policy Adaptation using Latent Space Search

Felix Chalumeau, Shikha Surana, Clément Bonnet, Nathan Grinsztajn, Arnu Pretorius, Alexandre Laterre +1

NeurIPS 2023poster

Combinatorial Optimization underpins many real-world applications and yet, designing performant algorithms to solve these complex, typically NP-hard, problems remains a significant research challenge. Reinforcement Learning (RL) provides a versatile framework for designing heuristics across a broad…

Cited by 27 Source ↗PDF ↗code ↗Scholar ↗

2023

Nonparametric Boundary Geometry in Physics Informed Deep Learning

Scott Alexander Cameron, Arnu Pretorius, Stephen J. Roberts

NeurIPS 2023poster

Engineering design problems frequently require solving systems of partial differential equations with boundary conditions specified on object geometries in the form of a triangular mesh. These boundary geometries are provided by a designer and are problem dependent. The efficiency of the design proc…

Cited by 2 Source ↗PDF ↗Scholar ↗

2022

Robust and Scalable SDE Learning: A Functional Perspective

Scott Alexander Cameron, Tyron Luke Cameron, Arnu Pretorius, Stephen J. Roberts

ICLR 2022poster

Stochastic differential equations provide a rich class of flexible generative models, capable of describing a wide range of spatio-temporal processes. A host of recent work looks to learn data-representing SDEs, using neural networks and other flexible function approximators. Despite these advances,…

Cited by 3 Source ↗PDF ↗Scholar ↗

2022

Towards a Standardised Performance Evaluation Protocol for Cooperative MARL

Rihab Gorsane, Omayma Mahjoub, Ruan John de Kock, Roland Dubb, Siddarth Singh, Arnu Pretorius

NeurIPS 2022accept

Multi-agent reinforcement learning (MARL) has emerged as a useful approach to solving decentralised decision-making problems at scale. Research in the field has been growing steadily with many breakthrough algorithms proposed in recent years. In this work, we take a closer look at this rapid develop…

Cited by 56 Source ↗PDF ↗Scholar ↗

2022

Universally Expressive Communication in Multi-Agent Reinforcement Learning

Matthew Morris, Thomas D Barrett, Arnu Pretorius

NeurIPS 2022accept

Allowing agents to share information through communication is crucial for solving complex tasks in multi-agent reinforcement learning. In this work, we consider the question of whether a given communication protocol can express an arbitrary policy. By observing that many existing protocols can be vi…

Cited by 8 Source ↗PDF ↗code ↗Scholar ↗

2020

A game-theoretic analysis of networked system control for common-pool resource management using multi-agent reinforcement learning

Arnu Pretorius, Scott Cameron, Elan van Biljon, Thomas Makkink, Shahil Mawjee, Jeremy du Plessis +3

NeurIPS 2020poster

Multi-agent reinforcement learning has recently shown great promise as an approach to networked system control. Arguably, one of the most difficult and important tasks for which large scale networked system control is applicable is common-pool resource management. Crucial common-pool resources inclu…

Cited by 15 Source ↗PDF ↗code ↗Scholar ↗

2018

Critical initialisation for deep signal propagation in noisy rectifier neural networks

Arnu Pretorius, Elan van Biljon, Steve Kroon, Herman Kamper

NeurIPS 2018poster

Stochastic regularisation is an important weapon in the arsenal of a deep learning practitioner. However, despite recent theoretical advances, our understanding of how noise influences signal propagation in deep neural networks remains limited. By extending recent work based on mean field theory, we…

Cited by 20 Source ↗PDF ↗code ↗Scholar ↗

2018

Learning Dynamics of Linear Denoising Autoencoders

Arnu Pretorius, Steve Kroon, Herman Kamper

ICML 2018oral

Denoising autoencoders (DAEs) have proven useful for unsupervised representation learning, but a thorough theoretical understanding is still lacking of how the input noise influences learning. Here we develop theory for how noise influences learning in DAEs. By focusing on linear DAEs, we are able t…

Cited by 38 Source ↗PDF ↗code ↗Scholar ↗