Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Hallucinative Topological Memory for Zero-Shot Visual Planning
Kara Liu, Thanard Kurutach, Christine Tung, Pieter Abbeel, Aviv Tamar
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 |
|---|---|---|---|
| Use HTM instead of Visual Foresight with green pixel distance | Visual planning on block wall · Block wall (Domain 1) | Execution success: 90% -> 95% | Improved |
| Use HTM instead of SPTM with CVAE | Visual planning on block wall with complex obstacle · Block wall with complex obstacle (Domain 2) | Execution success: 30% -> 100% | Improved |
| Use HTM instead of Visual Foresight | Block insertion planning · Block insertion (hard) | Execution success: 10 % -> 70 % | Improved |
| Replace the BCE connectivity classifier with a CPC energy model (ablation) | SPTM connectivity scoring · Block wall domains | Plan feasibility | Improved |
| Use inverse score-based edge weights instead of thresholded binary edges (ablation) | CPC connectivity graph · Block wall with complex obstacle | Average distance to goal | Improved |
| Add negative samples generated from the conditional model (ablation) | CPC connectivity model · Hallucinated images | Connectivity-score consistency | Improved |
| Plan from hallucinated context-conditioned images rather than requiring exploration samples in the new environment | HTM · Unseen environment contexts | Zero-shot planning | Improved |
Closest papers by meaning
- Pose Invariant Topological Memory for Visual Navigation (ICCV 2021 · similarity 0.90)
- Zero-Shot Robotic Manipulation with Pre-Trained Image-Editing Diffusion Models (ICLR 2024 · similarity 0.86)
- Embodied Task Planning via Graph-Informed Action Generation with Large Language Models (ICML 2026 · similarity 0.86)
- Differentiable Learning of Lifted Action Schemas for Classical Planning (NeurIPS 2026 · similarity 0.86)
- Visual Planning: Let's Think Only with Images (ICLR 2026 · similarity 0.86)
- Reasoning in Visual Navigation of End-to-end Trained Agents: A Dynamical Systems Approach (CVPR 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.
