Research map / Reinforcement learning / MDPs and planning
Unveiling Options with Neural Network Decomposition
Read the paperCheck your idea against 108,967 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.
Decomposes neural network policies for MDPs into reusable sub-policies and synthesizes them into temporally extended actions (options).
Problem
RL agents learn task-specific policies that do not generalize to related tasks; reusable temporally extended behaviors are needed.
Approach
Decompose the neural network encoding a policy into reusable sub-policies and compose them into options, rather than learning options from scratch.
Closest papers by meaning
- Reward-Respecting Subtasks for Model-Based Reinforcement Learning (Abstract Reprint) (AAAI 2024 · similarity 0.88)
- Efficient Reinforcement Learning by Discovering Neural Pathways (NeurIPS 2024 · similarity 0.88)
- Mitigating Information Loss in Tree-Based Reinforcement Learning via Direct Optimization (ICLR 2025 · similarity 0.87)
- In Search of Trees: Decision-Tree Policy Synthesis for Black-Box Systems via Search (AAAI 2025 · similarity 0.87)
- Autonomous Option Invention for Continual Hierarchical Reinforcement Learning and Planning (AAAI 2025 · similarity 0.86)
- SPOT: Scalable Policy Optimization with Trees for Markov Decision Processes (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.