67 citations · 136 across the 17 of their papers we have counts for
17 papers
DiffusionGemma Technical Report
DiffusionGemma Team, Adrien Ali Taïga, James Assiene +41
We introduce DiffusionGemma, an experimental open-weight language model that uses discrete diffusion to generate text at exceptionally high speed. Rather than decoding one token at…
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…
Acceleration in Policy Optimization
Veronica Chelu, Tom Zahavy, Arthur Guez +2
We work towards a unifying paradigm for accelerating policy optimization methods in reinforcement learning (RL) by integrating foresight in the policy improvement step via optimist…
Discovering Attention-Based Genetic Algorithms via Meta-Black-Box Optimization
Robert Tjarko Lange, Tom Schaul, Yutian Chen +4
Genetic algorithms constitute a family of black-box optimization algorithms, which take inspiration from the principles of biological evolution. While they provide a general-purpos…
ReLOAD: Reinforcement Learning with Optimistic Ascent-Descent for Last-Iterate Convergence in Constrained MDPs
Ted Moskovitz, Brendan O'Donoghue, Vivek Veeriah +3
In recent years, Reinforcement Learning (RL) has been applied to real-world problems with increasing success. Such applications often require to put constraints on the agent's beha…
Optimistic Meta-Gradients
Sebastian Flennerhag, Tom Zahavy, Brendan O'Donoghue +3
We study the connection between gradient-based meta-learning and convex op-timisation. We observe that gradient descent with momentum is a special case of meta-gradients, and build…
Discovering Evolution Strategies via Meta-Black-Box Optimization
Robert Tjarko Lange, Tom Schaul, Yutian Chen +5
Optimizing functions without access to gradients is the remit of black-box methods such as evolution strategies. While highly general, their learning dynamics are often times heuri…
Probing Transfer in Deep Reinforcement Learning without Task Engineering
Andrei A. Rusu, Sebastian Flennerhag, Dushyant Rao +2
We evaluate the use of original game curricula supported by the Atari 2600 console as a heterogeneous transfer benchmark for deep reinforcement learning agents. Game designers crea…
Meta-Gradients in Non-Stationary Environments
Jelena Luketina, Sebastian Flennerhag, Yannick Schroecker +3
Meta-gradient methods (Xu et al., 2018; Zahavy et al., 2020) offer a promising solution to the problem of hyperparameter selection and adaptation in non-stationary reinforcement le…
Discovering Policies with DOMiNO: Diversity Optimization Maintaining Near Optimality
Tom Zahavy, Yannick Schroecker, Feryal Behbahani +4
Finding different solutions to the same problem is a key aspect of intelligence associated with creativity and adaptation to novel situations. In reinforcement learning, a set of d…
Introducing Symmetries to Black Box Meta Reinforcement Learning
Louis Kirsch, Sebastian Flennerhag, Hado van Hasselt +3
Meta reinforcement learning (RL) attempts to discover new RL algorithms automatically from environment interaction. In so-called black-box approaches, the policy and the learning a…
Bootstrapped Meta-Learning
Sebastian Flennerhag, Yannick Schroecker, Tom Zahavy +3
Meta-learning empowers artificial intelligence to increase its efficiency by learning how to learn. Unlocking this potential involves overcoming a challenging meta-optimisation pro…
Discovering Diverse Nearly Optimal Policies with Successor Features
Tom Zahavy, Brendan O'Donoghue, Andre Barreto +3
Finding different solutions to the same problem is a key aspect of intelligence associated with creativity and adaptation to novel situations. In reinforcement learning, a set of d…
Temporal Difference Uncertainties as a Signal for Exploration
Sebastian Flennerhag, Jane X. Wang, Pablo Sprechmann +7
An effective approach to exploration in reinforcement learning is to rely on an agent's uncertainty over the optimal policy, which can yield near-optimal exploration strategies in…
Meta-Learning with Warped Gradient Descent
Sebastian Flennerhag, Andrei A. Rusu, Razvan Pascanu +3
Learning an efficient update rule from data that promotes rapid learning of new tasks from the same distribution remains an open problem in meta-learning. Typically, previous works…
Transferring Knowledge across Learning Processes
Sebastian Flennerhag, Pablo G. Moreno, Neil D. Lawrence +1
In complex transfer learning scenarios new tasks might not be tightly linked to previous tasks. Approaches that transfer information contained only in the final parameters of a sou…
Breaking the Activation Function Bottleneck through Adaptive Parameterization
Sebastian Flennerhag, Hujun Yin, John Keane +1
Standard neural network architectures are non-linear only by virtue of a simple element-wise activation function, making them both brittle and excessively large. In this paper, we…