Research map / Reinforcement learning / Visual and representation RL
Relational Reinforcement Learning for Planning with Exogenous Effects
David Martínez, Guillem Alenyà, Tony Ribeiro, Katsumi Inoue, Carme Torras
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add the confidence term to the operator score function (ablation) | Proposed model learner · Triangle Tireworld MDP-2 | variational distance | Improved |
| vary the penalty scale alpha (0.005, 0.02, 0.04) (ablation) | Proposed model learner · Triangle Tireworld MDP-2 | variational distance | No clear effect |
| increase candidate set limit kappa from 500 to 2000 (ablation) | Proposed model learner · Triangle Tireworld MDP-2 | variational distance | No clear effect |
| increase candidate set limit kappa (ablation) | Proposed model learner · Crossing Traffic MDP-1 | variational distance and learning time | Mixed |
| use non-admissible heuristic (delta = 0.1 instead of 0) (ablation) | Proposed model learner · Crossing Traffic MDP-1 | search speed | Improved |
| add the OI-subsumption tree (ablation) | Proposed model learner · Crossing Traffic MDP-1 | variational distance and learning time | Improved |
| replace Pasula et al. learner with proposed learner | Model learning · Crossing Traffic MDP-1 and Elevators MDP-1 | variational distance | Improved |
| replace Pasula et al. learner with proposed learner | Model learning · Triangle Tireworld MDP-2 | variational distance | No clear effect |
| use proposed learner instead of Pasula et al. inside V-MIN | V-MIN · Triangle Tireworld MDP-4 | total teacher requests: 10.77 -> 8.34 | Improved |
| use proposed learner instead of Pasula et al. inside V-MIN | V-MIN · Triangle Tireworld MDP-2 | total teacher requests: 6.31 -> 6.48 | No clear effect |
| add teacher demonstrations (V-MIN instead of REX) | REX with proposed learner · Crossing Traffic MDP-1 to MDP-3 | accumulated reward | Improved |
Closest papers by meaning
- ExoPredicator: Learning Abstract Models of Dynamic Worlds for Robot Planning (ICLR 2026 · similarity 0.87)
- Plan-Seq-Learn: Language Model Guided RL for Solving Long Horizon Robotics Tasks (ICLR 2024 · similarity 0.86)
- Learning to Execute: Efficient Learning of Universal Plan-Conditioned Policies in Robotics (NeurIPS 2021 · similarity 0.86)
- GLIB: Efficient Exploration for Relational Model-Based Reinforcement Learning via Goal-Literal Babbling (AAAI 2021 · similarity 0.86)
- Hierarchical Planning and Learning for Robots in Stochastic Settings Using Zero-Shot Option Invention (AAAI 2024 · similarity 0.86)
- Efficient Reinforcement Learning in Probabilistic Reward Machines (AAAI 2025 · similarity 0.84)
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.
