Research map / Vision-language and multimodal / Robot action and navigation
Prototypical Action Reasoning Facilitated by Vision-Language Alignment for Egocentric Action Anticipation
Jiang Shao, Xinbo Zhao, Wenyin Tuo, Xiaochun Zou
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 prototypical action reasoning (PAR) branch (ablation) | PAR-VLA (TSF-B) · EPIC-KITCHENS-100 val | action Top-5 recall: 19.6 -> 21.7 | Improved |
| add vision-language alignment (VLA) supervision (ablation) | PAR-VLA (TSF-B) · EPIC-KITCHENS-100 val | action Top-5 recall: 19.6 -> 20.3 | Improved |
| combine PAR and VLA (ablation) | PAR-VLA (TSF-B) · EPIC-KITCHENS-100 val | action Top-5 recall: 19.6 -> 22.5 | Improved |
| add PAR branch (ablation) | PAR-VLA (TSF-B) · EGTEA | Top-5 recall: 66.5 -> 68.4 | Improved |
| add VLA supervision (ablation) | PAR-VLA (TSF-B) · EGTEA | Top-5 recall: 66.5 -> 68.1 | Improved |
| combine PAR and VLA (ablation) | PAR-VLA (TSF-B) · EGTEA | Top-5 recall: 66.5 -> 70.2 | Improved |
| add optical flow modality to RGB input | PAR-VLA · EPIC-KITCHENS-100 val | action Top-5 recall: 22.5 -> 24.1 | Improved |
| set prototype set size K per category to 5 (ablation) | PAR-VLA · EPIC-KITCHENS-100 val | Top-5 recall | Improved |
| stack N_par = 6 PCR encoding layers (ablation) | PAR-VLA · EPIC-KITCHENS-100 val | Top-5 recall | Improved |
| PAR-VLA versus SGEAR-2B on 50-Salads | PAR-VLA · 50-Salads | Top-1 accuracy: 63.9 -> 65.4 | Improved |
Closest papers by meaning
- ATI-VLA: Action-Centric Predictive Vision–Language–Action Models via Actionable Alignment Then Adaptive Injection (NeurIPS 2026 · similarity 0.89)
- Can't Make an Omelette Without Breaking Some Eggs: Plausible Action Anticipation Using Large Video-Language Models (CVPR 2024 · similarity 0.89)
- Summarize the Past to Predict the Future: Natural Language Descriptions of Context Boost Multimodal Object Interaction Anticipation (CVPR 2024 · similarity 0.88)
- DreamVLA: A Vision-Language-Action Model Dreamed with Comprehensive World Knowledge (NeurIPS 2025 · similarity 0.88)
- Uncertainty-aware Action Decoupling Transformer for Action Anticipation (CVPR 2024 · similarity 0.88)
- UP-VLA: A Unified Understanding and Prediction Model for Embodied Agent (ICML 2025 · similarity 0.88)
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.
