Research map / Vision-language and multimodal / Cross-modal learning
Cut to the Chase: Training-free Multimodal Summarization via Chain-of-Events
Xiaoxing You, Qiang Huang, Lingyu Li, Xiaojun Chang, Jun Yu
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 HEG construction (ablation) | CoE · XMSMO | CIDEr: 44.09 -> 56.55 | Improved |
| add HEG construction (ablation) | CoE · MM-AVS | ROUGE: 16.68 -> 19.53 | Improved |
| add CSG grounding (ablation) | CoE · MM-AVS | ROUGE: 16.35 -> 19.53 | Improved |
| add CSG grounding (ablation) | CoE · VIEWS | ROUGE: 18.56 -> 19.28 | Improved |
| add DSG style adaptation (ablation) | CoE · VIEWS | CIDEr: 9.23 -> 19.73 | Improved |
| add DSG style adaptation (ablation) | CoE · XMSMO | CIDEr: 40.56 -> 56.55 | Improved |
| add EER event evolution reasoning (ablation) | CoE · MM-AVS | CIDEr: 13.83 -> 17.89 | Improved |
| add HEG construction (ablation) | CoE · SummScreen3D | ROUGE: 16.25 vs 16.15 | Hurt |
| add HEG construction (ablation) | CoE · BLiSS | CIDEr: 8.02 vs 7.19 | Hurt |
| add DSG style adaptation (ablation) | CoE · VIEWS | ROUGE: 18.76 -> 19.28 | No clear effect |
| use CoE framework over base MLLM backbones | CoE · 8 MMS datasets | ROUGE: +1.0 on VISTA to +5.1 on XMSMO | Improved |
| scale backbone from 3B to 7B | CoE · Summ and VIEWS | ROUGE: 3B slightly surpasses 7B | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TCoT — TCoT: Trajectory Chain-of-Thoughts for Robotic Manipulation with Failure Recovery in Vision-Language-Action Model (AAAI 2026) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MMSum: A Dataset for Multimodal Summarization and Thumbnail Generation of Videos (CVPR 2024 · similarity 0.88)
- V2Xum-LLM: Cross-Modal Video Summarization with Temporal Prompt Instruction Tuning (AAAI 2025 · similarity 0.87)
- Beyond Grounding: Extracting Fine-Grained Event Hierarchies across Modalities (AAAI 2024 · similarity 0.87)
- EventLens: Event-Structure Reinforcement Learning for Video Understanding (NeurIPS 2026 · similarity 0.87)
- SF2T: Self-supervised Fragment Finetuning of Video-LLMs for Fine-Grained Understanding (CVPR 2025 · similarity 0.87)
- Video Summarization with Large Language Models (CVPR 2025 · 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.
