Research map / Graph learning / Graph algorithms and clustering
Learning to Plan with Uncertain Topological Maps
Edward Beeching, Jilles Dibangoye, Olivier Simonin, Christian Wolf
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 visual node features to the neural planner (ablation) | neural graph planner · uncertain graphs | H-SPL: 0.468 -> 0.501 | Improved |
| add visual node features to the neural planner (ablation) | neural graph planner · uncertain graphs | accuracy: 0.251 -> 0.262 | Improved |
| use neural planner without visual features instead of best symbolic baseline | neural graph planner · uncertain graphs | H-SPL: 0.269 -> 0.468 | Improved |
| add modality mixing (mod-drop with GT adjacency) during training (ablation) | neural graph planner · uncertain graphs | accuracy: 2.3% improvement | Improved |
| sample point-goals from planner output instead of argmax | neural planner + local policy · hierarchical navigation | SPL: 0.877 -> 0.796 | Hurt |
| sample point-goals from planner output instead of argmax | neural planner + local policy · hierarchical navigation | success rate: 0.983 -> 0.966 | Hurt |
| use neural planner instead of symbolic custom cost | neural planner + local policy · hierarchical navigation | success rate: 0.707 -> 0.983 | Improved |
| add GRU for message accumulation (Bellman-Ford bias) to graph network (ablation) | neural graph planner · planner accuracy vs data size 8,000 to 74,000 graphs | accuracy | Improved |
| increase local policy steps m up to 20 time-steps (ablation) | hierarchical agent · 1,200 validation instances | success: 20 time-steps | No clear effect |
Closest papers by meaning
- Hallucinative Topological Memory for Zero-Shot Visual Planning (ICML 2020 · similarity 0.88)
- Graph Learning for Numeric Planning (NeurIPS 2024 · similarity 0.86)
- A Bayesian Approach to Online Planning (ICML 2024 · similarity 0.85)
- Learning Neuro-Symbolic Abstractions for Robot Planning and Learning (AAAI 2024 · similarity 0.85)
- Learning Cognitive Maps from Transformer Representations for Efficient Planning in Partially Observed Environments (ICML 2024 · similarity 0.85)
- State Encodings for GNN-Based Lifted Planners (AAAI 2025 · similarity 0.85)
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.
