Research map / Vision-language and multimodal / Robot action and navigation
ESCAPE: Episodic Spatial Memory and Adaptive Execution Policy for Long-Horizon Mobile Manipulation
Jingjing Qian, Zeyuan He, Chen Shi, Lei Xiao, Li Jiang
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 Memory Retrieval and Update (temporal memory) (ablation) | ESCAPE · ALFRED valid unseen | SR: 27.16 -> 61.27 | Improved |
| add Observation-to-Memory Encoding (ablation) | ESCAPE · ALFRED valid seen | SR: 53.54 -> 62.32 | Improved |
| add Adaptive Execution Policy (reactive local monitor) (ablation) | ESCAPE · ALFRED valid seen | SR: 57.59 -> 62.32 | Improved |
| add Adaptive Execution Policy (ablation) | ESCAPE · ALFRED valid seen | PLWSR: 38.44 -> 44.24 | Improved |
| add Adaptive Execution Policy (ablation) | ESCAPE · ALFRED valid seen | efficiency factor: 0.668 -> 0.710 | Improved |
| use 100x100 grid instead of 80x80 (ablation) | ESCAPE · ALFRED valid seen | SR: 52.68 -> 62.32 | Improved |
| use 100x100 grid instead of 120x120 (ablation) | ESCAPE · ALFRED valid seen | SR: 53.54 -> 62.32 | Improved |
| use memory-derived dynamic object queries instead of static queries (ablation) | ESCAPE · ALFRED valid unseen | SR: 58.83 -> 61.27 | Improved |
| replace predicted masks with ground-truth simulator masks (ablation) | ESCAPE · ALFRED valid seen | SR: 62.32 -> 62.56 | No clear effect |
| replace predicted masks with ground-truth simulator masks (ablation) | ESCAPE · ALFRED valid unseen | SR: 61.27 -> 63.82 | Improved |
| ESCAPE versus DISCO without step-by-step instructions, unseen GC | ESCAPE · ALFRED test unseen (goal only) | GC: 65.50 -> 64.79 | Hurt |
| ESCAPE versus DISCO path-length-weighted success | ESCAPE · ALFRED test seen (step-by-step instructions) | PLWSR: 40.60 -> 52.42 | Improved |
Closest papers by meaning
- Planning with an Embodied Learnable Memory (ICLR 2026 · similarity 0.87)
- EchoVLA: Robotic Vision-Language-Action Model with Synergistic Declarative Memory for Mobile Manipulation (ECCV 2026 · similarity 0.87)
- SAM2Act: Integrating Visual Foundation Model with A Memory Architecture for Robotic Manipulation (ICML 2025 · similarity 0.87)
- ManiSkill-HAB: A Benchmark for Low-Level Manipulation in Home Rearrangement Tasks (ICLR 2025 · similarity 0.86)
- AC-DiT: Adaptive Coordination Diffusion Transformer for Mobile Manipulation (NeurIPS 2025 · similarity 0.86)
- SPIN: Simultaneous Perception Interaction and Navigation (CVPR 2024 · 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.
