Research map / Video / Video-language understanding
Non-Autoregressive Coarse-to-Fine Video Captioning
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 visual word generation loss (ablation) | NA-B · MSR-VTT | CIDEr-D: 47.6 -> 49.4 | Helped |
| Add visual word generation loss (ablation) | AR-B · MSR-VTT | CIDEr-D: 49.1 -> 50.8 | Helped |
| Add coarse-grained visual-word templates (ablation) | NA-B with visual word generation loss · MSR-VTT | CIDEr-D: 49.4 -> 51.4 | Helped |
| Use coarse-grained templates with MP rather than MP from scratch (ablation) | NACF · MSR-VTT | CIDEr-D: 49.4 -> 51.4 | Helped |
| Use coarse-grained templates with EF rather than EF from scratch (ablation) | NACF · MSR-VTT | CIDEr-D: 50.6 -> 51.8 | Helped |
| Use coarse-grained templates with L2R rather than L2R from scratch (ablation) | NACF · MSR-VTT | CIDEr-D: 48.7 -> 50.6 | Helped |
| Replace monotonic L2R decoding with adaptive EF decoding (ablation) | NACF without templates · MSR-VTT | CIDEr-D: 48.7 -> 50.6 | Helped |
| Replace monotonic CT-L2R decoding with adaptive CT-EF decoding (ablation) | NACF with templates · MSR-VTT | CIDEr-D: 50.6 -> 51.8 | Helped |
| Use NACF instead of AR-B | Video captioning · MSVD | METEOR: 36.9 -> 36.2 | Hurt |
| Use NACF rather than AR-B for parallel caption decoding | Video captioning · MSVD | inference speedup: 3.2 times | Helped |
| Use NACF rather than AR-B for parallel caption decoding | Video captioning · MSR-VTT | inference speedup: 2.2 times | Helped |
| Use NACF rather than NA-B | Video captioning · MSR-VTT | CIDEr-D: 47.6 -> 51.4 | Helped |
Closest papers by meaning
- Fine-Grained Captioning of Long Videos through Scene Graph Consolidation (ICML 2025 · similarity 0.90)
- Distilling Vision-Language Models on Millions of Videos (CVPR 2024 · similarity 0.89)
- AuroraCap: Efficient, Performant Video Detailed Captioning and a New Benchmark (ICLR 2025 · similarity 0.88)
- Progress-Aware Video Frame Captioning (CVPR 2025 · similarity 0.88)
- Streaming Dense Video Captioning (CVPR 2024 · similarity 0.87)
- MM-SeR: Multimodal Self-Refinement for Lightweight Image Captioning (CVPR 2026 · similarity 0.87)
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.