Research map / Video / Action recognition
General Covariant Action Modeling: Constructing Generalized Manifolds via Spatio-Temporal Decoupling
Huaihai Lyu, Chaofan Chen, Mingyu Cao, Yuheng Ji, Changsheng Xu
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 SAF (ablation) | GAM (without ALP) · LIBERO-Long | success rate: 82.4 -> 89.1 | Improved |
| add ALP (ablation) | GAM (without SAF) · LIBERO-Long | success rate: 82.4 -> 76.4 | Hurt |
| add ALP to SAF (ablation) | GAM · LIBERO-Long | success rate: 89.1 -> 92.0 | Improved |
| add SAF to ALP (ablation) | GAM · LIBERO-Spatial | success rate: 91.6 -> 98.6 | Improved |
| add ALP to SAF (ablation) | GAM · LIBERO-Goal | success rate: 96.8 -> 96.3 | Hurt |
| add ALP to SAF (ablation) | GAM · LIBERO-Object | success rate: 97.3 -> 97.5 | No clear effect |
| add planner loss L_plan (ablation) | GAM · LIBERO-Long | success rate: 5.1 -> 92.0 | Improved |
| add velocity loss L_vel (ablation) | GAM · LIBERO-Long | success rate: 87.3 -> 92.0 | Improved |
| add velocity loss L_vel (ablation) | GAM · LIBERO-Spatial | success rate: 93.6 -> 98.6 | Improved |
| add ALP under speed noise delta=1.0 (ablation) | GAM · SimplerEnv Put Spoon | success rate: 12.5% -> 33.3% | Improved |
| add ALP under speed noise delta=1.0 (ablation) | GAM · SimplerEnv Put Carrot | success rate: 16.7% -> 45.8% | Improved |
| add ALP under speed noise delta=1.0 (ablation) | GAM · SimplerEnv Stack Block | success rate: 4.2% -> 25.0% | Improved |
Closest papers by meaning
- LAST: Bridging Vision-Language and Action Manifolds via Gromov-Wasserstein Alignment (ICML 2026 · similarity 0.89)
- Neural Implicit Action Fields: From Discrete Waypoints to Continuous Functions for Vision-Language-Action Models (ICML 2026 · similarity 0.86)
- GAE: Unleashing Physical Potential of VLM with Generalizable Action Expert (ICML 2026 · similarity 0.86)
- World Models as Group Actions (NeurIPS 2026 · similarity 0.86)
- Beyond World-Frame Action Heads: Motion-Centric Action Frames for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.86)
- Bringing Motion Taxonomies to Continuous Domains via GPLVM on Hyperbolic manifolds (ICML 2024 · 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.
