Research map / Video / Video-language understanding
Saying the Unseen: Video Descriptions via Dialog Agents
Ye Zhu, Yu Wu, Yi Yang, Yan Yan
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 dialog between agents instead of no dialog | Q-BOT MM · AVSD | CIDEr: 14.2 -> 22.1 | Improved |
| use discriminative instead of generative internal dialog | QA-C · AVSD | CIDEr: 22.9 -> 27.1 | Improved |
| add attention modules (ablation) | QA-C(G) · AVSD | CIDEr: 20.4 -> 22.9 | Improved |
| add attention modules (ablation) | QA-C(D) · AVSD | CIDEr: 25.3 -> 27.1 | Improved |
| add audio-visual LSTM (ablation) | QA-C(G) · AVSD | CIDEr: +2.6 | Improved |
| add audio-visual LSTM (ablation) | QA-C(D) · AVSD | CIDEr: +3.4 | Improved |
| add internal selective mechanism with sparse annotations (ablation) | QA-C(D) · AVSD | CIDEr: 26.3 -> 27.1 | Improved |
| use MM attention instead of IM attention | QA-C(D) · AVSD | CIDEr: 26.1 -> 27.1 | Improved |
| keep QA pair order instead of shuffling (ablation) | QA-C(G) · AVSD | CIDEr: -2.9 | Improved |
| keep QA pair order instead of shuffling (ablation) | QA-C(D) · AVSD | CIDEr: -1.7 | Improved |
| add audio input for A-BOT (ablation) | QA-C(D) · AVSD | CIDEr: 26.6 -> 27.1 | Improved |
| add input descriptions for A-BOT (ablation) | QA-C(G) · AVSD | CIDEr: 20.1 -> 22.9 | Improved |
Closest papers by meaning
- Describing Unseen Videos via Multi-Modal Cooperative Dialog Agents (ECCV 2020 · similarity 0.94)
- V^2Dial: Unification of Video and Visual Dialog via Multimodal Experts (CVPR 2025 · similarity 0.86)
- MMDU: A Multi-Turn Multi-Image Dialog Understanding Benchmark and Instruction-Tuning Dataset for LVLMs (NeurIPS 2024 · similarity 0.85)
- Visual Graph Arena: Evaluating Visual Conceptualization of Vision and Multimodal Large Language Models (ICML 2025 · similarity 0.85)
- Divide-and-Conquer: Tree-structured Strategy with Answer Distribution Estimator for Goal-Oriented Visual Dialogue (AAAI 2025 · similarity 0.85)
- PerceptionLM: Open-Access Data and Models for Detailed Visual Understanding (NeurIPS 2025 · 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.
