Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Representation-driven Option Discovery in Reinforcement Learning
Marlos C. Machado
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Presents a talk introducing a general framework that discovers RL options from the agent's own representation.
Problem
Modelling temporal abstraction in reinforcement learning via temporally extended actions (options), where useful options must be discovered automatically.
Approach
Uses the agent's learned representation to discover useful options, and then leverages these options to generate a rich set of behaviours.
Closest papers by meaning
- Temporal Abstraction in Reinforcement Learning with the Successor Representation (JMLR 2023 · similarity 0.93)
- Autonomous Option Invention for Continual Hierarchical Reinforcement Learning and Planning (AAAI 2025 · similarity 0.88)
- Reward-Respecting Subtasks for Model-Based Reinforcement Learning (Abstract Reprint) (AAAI 2024 · similarity 0.87)
- OptionZero: Planning with Learned Options (ICLR 2025 · similarity 0.86)
- Learning Generalizable and Composable Abstractions for Transfer in Reinforcement Learning (AAAI 2024 · similarity 0.86)
- Reward-Aware Proto-Representations in Reinforcement Learning (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
