Research map / Video / Action recognition
Leveraging Temporal Context in Low Representational Power Regimes
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 ETM supervision | MoViNet A0 · EPIC-KITCHENS-100 | action top-1 accuracy: 36.8 -> 40.2 | Helped |
| Add ETM supervision | MoViNet A0 · EGO4D LTA | action top-1 accuracy: 21.1 -> 22.0 | Helped |
| Add ETM supervision | MoViNet A0 · EGTEA Gaze+ | action top-1 accuracy: 60.4 -> 62.5 | Helped |
| Add ETM supervision | small and larger action-recognition encoders · EPIC-KITCHENS-100 | action top-1 accuracy: 20.1 -> 32.4; 30.4 -> 30.7 | Mixed |
| Add past- and future-vector ETM supervision | MoViNet A0 · EPIC-KITCHENS-100 | action top-1 accuracy: 36.8 -> 40.2 | Helped |
| Add ETM supervision | anticipation encoders · EPIC-KITCHENS-100 | class-mean recall@5: 8.0 -> 9.1 | Helped |
| Add ETM supervision | anticipation encoders · EPIC-KITCHENS-100 | class-mean recall@5: 13.4 -> 13.5 | Helped |
| Preserve coherent event-sequence information instead of shuffling matrix entries (ablation) | MoViNet A0 with ETM training · EPIC-KITCHENS-100 | action top-1 accuracy: 36.3 -> 40.2 | Helped |
| Preserve the correspondence between ETM rows and columns and action labels (ablation) | MoViNet A0 with ETM training · EPIC-KITCHENS-100 | action top-1 accuracy: 36.7 -> 40.2 | Helped |
| Use ordered temporal transitions rather than symmetric co-occurrence (ablation) | MoViNet A0 with ETM training · EPIC-KITCHENS-100 | action top-1 accuracy: 37.9 -> 40.2 | Helped |
| Add future-vector supervision alongside past-vector supervision (ablation) | MoViNet A0 with ETM training · EPIC-KITCHENS-100 | action top-1 accuracy: 38.2 -> 40.2 | Helped |
| Add past-vector supervision alongside future-vector supervision (ablation) | MoViNet A0 with ETM training · EPIC-KITCHENS-100 | action top-1 accuracy: 38.3 -> 40.2 | Helped |
Closest papers by meaning
- Energy vs. Noise: Towards Robust Temporal Action Localization in Open-World (AAAI 2025 · similarity 0.87)
- Summarize the Past to Predict the Future: Natural Language Descriptions of Context Boost Multimodal Object Interaction Anticipation (CVPR 2024 · similarity 0.87)
- TIM: A Time Interval Machine for Audio-Visual Action Recognition (CVPR 2024 · similarity 0.86)
- Chirality in Action: Time-Aware Video Representation Learning by Latent Straightening (NeurIPS 2025 · similarity 0.86)
- PASS: Path-selective State Space Model for Event-based Recognition (NeurIPS 2025 · similarity 0.86)
- ActFusion: a Unified Diffusion Model for Action Segmentation and Anticipation (NeurIPS 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.