Reinforcement Learning

PyCRM: A Python library for reward machine-based reinforcement learning
Reinforcement Learning (RL) research often models environments as Markov decision processes. Yet many real-world tasks are …
PDF Cite Code DOI
Drowning in Degrees of Freedom: One Agent for Every Task
We describe a research programme aimed at the construction of a single, generally intelligent agent—one competent across all …
PDF Cite Project
Redistribution-based Cost Inference Improves Sparse Safe Offline RL
Safe offline RL typically assumes access to dense per-step cost annotations, but in practice supervisors provide only trajectory-level …
PDF Cite
The Goal-Directed Frame for General Agents
Reinforcement learning is often framed around episodic, discounted, or average scalar rewards. While useful, these views miss a core …
PDF Cite
An Unreasonably Simple Approach to Safe RL
An important problem in reinforcement learning is designing agents that learn to solve tasks safely in an environment. A common …
PDF Cite Code
AI Agent Safety is a Reinforcement Learning Problem
With the rapid advancement and deployment of Agentic AI, our scientific understanding of capabilities and limitations has not kept …
PDF Cite
Weighted Composition for Entropy-regularised Reinforcement Learning
One avenue for creating a generally intelligent agent is to equip it with the ability to combine its previously learned behaviours to …
PDF Cite Project
MoralityGym: A Benchmark for Evaluating Hierarchical Moral Alignment in Sequential Decision-Making Agents
Evaluating moral alignment in agents navigating conflicting, hierarchically structured human norms is a critical challenge at the …
PDF Cite
Skill-Driven Neurosymbolic State Abstractions
We consider how to construct state abstractions compatible with a given set of abstract actions, to obtain a well-formed abstract …
PDF Cite Project
Finding the FrameStack: Learning What to Remember for Non-Markovian Reinforcement Learning
Recent success in developing increasingly general purpose agents based on sequence models has led to increased focus on the problem of …
PDF Cite