Research map / Video / Video-language understanding
MovieChat+: Question-Aware Sparse Memory for Long Video Question Answering
Enxin Song, Wenhao Chai, Ye Tian, Jenq–Neng Hwang, Xi Li, Gaoang Wang
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 question-aware memory consolidation (MovieChat+ over MovieChat) | MovieChat · MovieChat-1K global | accuracy: 62.3 -> 71.2 | Improved |
| add question-aware memory consolidation | MovieChat · MovieChat-1K breakpoint | accuracy: 48.3 -> 49.6 | Improved |
| add question-aware consolidation | MovieChat · MSVD-QA | accuracy: 75.2 -> 76.5 | No clear effect |
| add question-aware consolidation | MovieChat · EgoSchema | accuracy: 53.5 -> 56.4 | Improved |
| add memory mechanism (ablation) | MovieChat · MovieChat-1K global | accuracy: 51.4 -> 67.8 | Improved |
| add memory mechanism (ablation) | MovieChat · MovieChat-1K breakpoint | accuracy: 38.2 -> 50.4 | Improved |
| replace LLaMA with LLaMA2 decoder (ablation) | MovieChat · MovieChat-1K global | accuracy: 67.8 -> 64.2 | Hurt |
| replace LLaMA with LLaMA2 decoder (ablation) | MovieChat · MovieChat-1K breakpoint | accuracy: 50.4 -> 48.1 | Hurt |
| use more input frames with spatial pooling (ablation) | Spatial Pooling · ActivityNet CIDEr | CIDEr: 27.6 -> 27.4 | No clear effect |
| use more input frames with EMA (ablation) | EMA · ActivityNet CIDEr | CIDEr: 28.2 -> 16.3 | Hurt |
| use more frames with proposed memory (ablation) | MovieChat memory · ActivityNet CIDEr | CIDEr: 29.3 -> 29.3 | No clear effect |
| judge relevance by maximum rather than average similarity (ablation) | MovieChat+ · MovieChat-1K | accuracy | 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.
- MovieChat — MovieChat: From Dense Token to Sparse Memory for Long Video Understanding (CVPR 2024) · better on 9 of 9 shared results
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
- MovieChat: From Dense Token to Sparse Memory for Long Video Understanding (CVPR 2024 · similarity 0.95)
- Streaming Long Video Understanding with Large Language Models (NeurIPS 2024 · similarity 0.93)
- LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding (NeurIPS 2024 · similarity 0.91)
- ReWind: Understanding Long Videos with Instructed Learnable Memory (CVPR 2025 · similarity 0.91)
- TimeChat: A Time-sensitive Multimodal Large Language Model for Long Video Understanding (CVPR 2024 · similarity 0.91)
- Not Another Text Benchmark: Putting the “Visual" Back in Visual Question Answering for Large Video Models (NeurIPS 2026 · similarity 0.91)
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.
