21 citations · 135 across the 39 of their papers we have counts for
40 papers
Imagined Autocurricula
Ahmet H. Güzel, Matthew Thomas Jackson, Jarek Luca Liesen +4
Training agents to act in embodied environments typically requires vast training data or access to accurate simulation, neither of which exists for many cases in the real world. In…
Learning When to Plan: Efficiently Allocating Test-Time Compute for LLM Agents
Davide Paglieri, Bartłomiej Cupiał, Jonathan Cook +6
Training large language models (LLMs) to reason via reinforcement learning (RL) significantly improves their problem-solving capabilities. In agentic settings, existing methods lik…
Programming by Backprop: An Instruction is Worth 100 Examples When Finetuning LLMs
Jonathan Cook, Silvia Sapora, Arash Ahmadian +4
Large language models (LLMs) are typically trained to acquire behaviours from demonstrations or experience, yet much of their training data is declarative: instructions, rules, and…
Investigating Non-Transitivity in LLM-as-a-Judge
Yi Xu, Laura Ruis, Tim Rocktäschel +1
Automatic evaluation methods based on large language models (LLMs) are emerging as the standard tool for assessing the instruction-following abilities of LLM-based agents. The most…
BALROG: Benchmarking Agentic LLM and VLM Reasoning On Games
Davide Paglieri, Bartłomiej Cupiał, Samuel Coward +10
Large Language Models (LLMs) and Vision Language Models (VLMs) possess extensive knowledge and exhibit promising reasoning abilities, however, they still struggle to perform well i…
Procedural Knowledge in Pretraining Drives Reasoning in Large Language Models
Laura Ruis, Maximilian Mozes, Juhan Bae +7
The capabilities and limitations of Large Language Models have been sketched out in great detail in recent years, providing an intriguing yet conflicting picture. On the one hand,…
Open-Endedness is Essential for Artificial Superhuman Intelligence
Edward Hughes, Michael Dennis, Jack Parker-Holder +5
In recent years there has been a tremendous surge in the general capabilities of AI systems, mainly fuelled by training foundation models on internetscale data. Nevertheless, the c…
Outliers and Calibration Sets have Diminishing Effect on Quantization of Modern LLMs
Davide Paglieri, Saurabh Dash, Tim Rocktäschel +1
Post-Training Quantization (PTQ) enhances the efficiency of Large Language Models (LLMs) by enabling faster operation and compatibility with more accessible hardware through reduce…
Genie: Generative Interactive Environments
Jake Bruce, Michael Dennis, Ashley Edwards +22
We introduce Genie, the first generative interactive environment trained in an unsupervised manner from unlabelled Internet videos. The model can be prompted to generate an endless…
Rainbow Teaming: Open-Ended Generation of Diverse Adversarial Prompts
Mikayel Samvelyan, Sharath Chandra Raparthy, Andrei Lupu +9
As large language models (LLMs) become increasingly prevalent across many real-world applications, understanding and enhancing their robustness to adversarial attacks is of paramou…
Debating with More Persuasive LLMs Leads to More Truthful Answers
Akbir Khan, John Hughes, Dan Valentine +7
Common methods for aligning large language models (LLMs) with desired behaviour heavily rely on human-labelled data. However, as models grow increasingly sophisticated, they will s…
Multi-Agent Diagnostics for Robustness via Illuminated Diversity
Mikayel Samvelyan, Davide Paglieri, Minqi Jiang +2
In the rapidly advancing field of multi-agent systems, ensuring robustness in unfamiliar and adversarial settings is crucial. Notwithstanding their outstanding performance in famil…
Leading the Pack: N-player Opponent Shaping
Alexandra Souly, Timon Willi, Akbir Khan +4
Reinforcement learning solutions have great success in the 2-player general sum setting. In this setting, the paradigm of Opponent Shaping (OS), in which agents account for the lea…
H-GAP: Humanoid Control with a Generalist Planner
Zhengyao Jiang, Yingchen Xu, Nolan Wagener +5
Humanoid control is an important research challenge offering avenues for integration into human-centric infrastructures and enabling physics-driven humanoid animations. The dauntin…
Scaling Opponent Shaping to High Dimensional Games
Akbir Khan, Timon Willi, Newton Kwan +5
In multi-agent settings with mixed incentives, methods developed for zero-sum games have been shown to lead to detrimental outcomes. To address this issue, opponent shaping (OS) me…
Vision-Language Models as a Source of Rewards
Kate Baumli, Satinder Baveja, Feryal Behbahani +24
Building generalist agents that can accomplish many goals in rich open-ended environments is one of the research frontiers for reinforcement learning. A key limiting factor for bui…
Mix-ME: Quality-Diversity for Multi-Agent Learning
Garðar Ingvarsson, Mikayel Samvelyan, Bryan Lim +3
In many real-world systems, such as adaptive robotics, achieving a single, optimised solution may be insufficient. Instead, a diverse set of high-performing solutions is often requ…
Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks
Samyak Jain, Robert Kirk, Ekdeep Singh Lubana +5
Fine-tuning large pre-trained models has become the de facto strategy for developing both task-specific and general-purpose machine learning systems, including developing models th…
minimax: Efficient Baselines for Autocurricula in JAX
Minqi Jiang, Michael Dennis, Edward Grefenstette +1
Unsupervised environment design (UED) is a form of automatic curriculum learning for training robust decision-making agents to zero-shot transfer into unseen environments. Such aut…
JaxMARL: Multi-Agent RL Environments and Algorithms in JAX
Alexander Rutherford, Benjamin Ellis, Matteo Gallici +18
Benchmarks are crucial in the development of machine learning algorithms, with available environments significantly influencing reinforcement learning (RL) research. Traditionally,…
Stabilizing Unsupervised Environment Design with a Learned Adversary
Ishita Mediratta, Minqi Jiang, Jack Parker-Holder +3
A key challenge in training generally-capable agents is the design of training tasks that facilitate broad generalization and robustness to environment variations. This challenge m…
MAESTRO: Open-Ended Environment Design for Multi-Agent Reinforcement Learning
Mikayel Samvelyan, Akbir Khan, Michael Dennis +5
Open-ended learning methods that automatically generate a curriculum of increasingly challenging tasks serve as a promising avenue toward generally capable reinforcement learning a…
General Intelligence Requires Rethinking Exploration
Minqi Jiang, Tim Rocktäschel, Edward Grefenstette
We are at the cusp of a transition from "learning from data" to "learning what data to learn from" as a central focus of artificial intelligence (AI) research. While the first-orde…
Learning General World Models in a Handful of Reward-Free Deployments
Yingchen Xu, Jack Parker-Holder, Aldo Pacchiano +5
Building generally capable agents is a grand challenge for deep reinforcement learning (RL). To approach this challenge practically, we outline two key desiderata: 1) to facilitate…