Research map / Reinforcement learning / Visual and representation RL
Induction of Subgoal Automata for Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| induce subgoal automata from traces rather than provide the automaton in advance | ISA vs QRM · OfficeWorld tasks | learning convergence | No clear effect |
| add reward shaping based on automaton-state distance (ablation) | ISA · OfficeWorld tasks | convergence speed | Helped |
| share experience across tasks (ablation) | ISA · OfficeWorld multitask tasks | convergence speed | Helped |
| reduce the observable set to only task-relevant observables (ablation) | ISA automaton learner · OfficeWorld CoffeeMail | automata computation time: 92% faster | Helped |
| reduce the observable set to only task-relevant observables (ablation) | ISA automaton learner · OfficeWorld CoffeeMail | number of examples: 46% smaller | Helped |
| use multitask training instead of single-task training (ablation) | ISA · OfficeWorld tasks | average ILASP running time | Mixed |
| add reward shaping (ablation) | ISA · OfficeWorld VisitABCD | average ILASP running time: 63.0 (11.4) -> 68.5 (13.0); 48.4 (8.8) -> 69.6 (8.1) | Hurt |
Closest papers by meaning
- Automating the Refinement of Reinforcement Learning Specifications (ICLR 2026 · similarity 0.88)
- ARM-FM: Automated Reward Machines via Foundation Models for Compositional Reinforcement Learning (ICLR 2026 · similarity 0.88)
- One Subgoal at a Time: Zero-Shot Generalization to Arbitrary Linear Temporal Logic Requirements in Multi-Task Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
- Meta Reinforcement Learning with Autonomous Inference of Subtask Dependencies (ICLR 2020 · similarity 0.87)
- DeepLTL: Learning to Efficiently Satisfy Complex LTL Specifications for Multi-Task RL (ICLR 2025 · similarity 0.87)
- Learning to Reason without External Rewards (ICLR 2026 · 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.