Research map / Video / Video-language understanding
Motion-Appearance Co-Memory Networks for Video Question Answering
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 |
|---|---|---|---|
| Replace fact concatenation with co-memory attention | Video QA network · TGIF-QA repetition action | ACC (%): 65.0 -> 66.8 | Helped |
| Replace memory concatenation with co-memory attention | Video QA network · TGIF-QA state transition | ACC (%): 70.7 -> 73.2 | Helped |
| Add multi-level contextual facts and dynamic fact ensemble (ablation) | Fact-concatenation memory network · TGIF-QA repetition action | ACC (%): 65.0 -> 66.3 | Helped |
| Add multi-level contextual facts and dynamic fact ensemble (ablation) | Fact-concatenation memory network · TGIF-QA repetition count | MSE: 4.34 -> 4.30 | Helped |
| Use two memory update cycles instead of one (ablation) | Co-memory attention model · TGIF-QA state transition | ACC (%): 69.9 -> 73.2 | Helped |
| Use three memory update cycles instead of two (ablation) | Co-memory attention model · TGIF-QA repetition action | ACC (%): 66.8 -> 66.5 | Hurt |
| Use three memory update cycles instead of two (ablation) | Co-memory attention model · TGIF-QA frame QA | ACC (%): 51.0 -> 51.1 | No clear effect |
| Add dynamic fact ensemble to co-memory attention with T=2 (ablation) | Co-memory network · TGIF-QA state transition | ACC (%): 73.2 -> 74.3 | Helped |
| Add co-memory attention using cross-cues from motion and appearance | Video QA network · TGIF-QA frame QA | ACC (%): 50.2 -> 51.0 | Helped |
Closest papers by meaning
- DynFocus: Dynamic Cooperative Network Empowers LLMs with Video Understanding (CVPR 2025 · similarity 0.87)
- Temporal Chain of Thought: Long-Video Understanding by Thinking in Frames (NeurIPS 2025 · similarity 0.87)
- Language-aware Visual Semantic Distillation for Video Question Answering (CVPR 2024 · similarity 0.87)
- STAIR: Spatial-Temporal Reasoning with Auditable Intermediate Results for Video Question Answering (AAAI 2024 · similarity 0.86)
- Question-Aware Gaussian Experts for Audio-Visual Question Answering (CVPR 2025 · similarity 0.86)
- Not Another Text Benchmark: Putting the “Visual" Back in Visual Question Answering for Large Video Models (NeurIPS 2026 · 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.