Research map / Video / Video-language understanding
Learning Video Context as Interleaved Multimodal Sequences
Kevin Qinghong Lin, Pengchuan Zhang, Difei Gao, Xide Xia, Joya Chen, Ziteng Gao, Jinheng Xie, Xuhong Xiao, Mike Zheng Shou
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 subtitles to video input (ablation) | MovieSeq · LVU relation | accuracy: 61.0 -> 75.6 | Improved |
| add subtitles to video input (ablation) | MovieSeq · LVU scene | accuracy: 76.8 -> 65.9 | Hurt |
| add subtitles to video input (ablation) | MovieSeq · CMD video-text retrieval | Geometric Mean: 9.2 -> 38.9 | Improved |
| add subtitles to video input (ablation) | MovieSeq · TVC captioning | CIDEr: 50.9 -> 63.5 | Improved |
| add character photos as context (ablation) | MovieSeq · TVC captioning | CIDEr: 63.5 -> 64.8 | No clear effect |
| add retrieved movie plot as context (ablation) | MovieSeq · MovieQA | accuracy: 27.32 -> 66.41 | Improved |
| add movie plot context (ablation) | MovieSeq · MADv2 | CIDEr: 11.4 -> 16.9 | Improved |
| add past clip narrations as context (ablation) | MovieSeq · MADv2 | Rep@4: 1.33 -> 0.11 | Improved |
| character identification via Movienet-trained module instead of CLIP retrieval (ablation) | MovieSeq · MADv2 | CIDEr: 19.7 -> 23.6 | Improved |
| multi-task co-training instead of individual training (ablation) | MovieSeq · MAD/MovieNet/TVC dev sets | average: 53.8 -> 55.2 | Improved |
| add subtitles on MovieQA video-only baseline (ablation) | MovieSeq · MovieQA | accuracy: 27.32 -> 48.32 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- MA-LMM — MA-LMM: Memory-Augmented Large Multimodal Model for Long-Term Video Understanding (CVPR 2024) · better on 1 of 1 shared result
- SwinBERT — SwinBERT: End-to-End Transformers With Sparse Attention for Video Captioning (CVPR 2022) · 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
- LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding (NeurIPS 2024 · similarity 0.89)
- CALVIN: Improved Contextual Video Captioning via Instruction Tuning (NeurIPS 2024 · similarity 0.88)
- InternVid: A Large-scale Video-Text Dataset for Multimodal Understanding and Generation (ICLR 2024 · similarity 0.88)
- Panda-70M: Captioning 70M Videos with Multiple Cross-Modality Teachers (CVPR 2024 · similarity 0.88)
- SAMA: Towards Multi-Turn Referential Grounded Video Chat with Large Language Models (NeurIPS 2025 · similarity 0.87)
- SF2T: Self-supervised Fragment Finetuning of Video-LLMs for Fine-Grained Understanding (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.
