Research map / Vision-language and multimodal / Visual question answering
Movie Question Answering: Remembering the Textual Cues for Layered Visual Contents
Bo Wang, Youjiang Xu, Yahong Han, Richang Hong
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 |
|---|---|---|---|
| Use LMN frame-level visual representations instead of MemN2N | MovieQA video-only · MovieQA Video | accuracy: 23.1 -> 38.6 | Improved |
| Use LMN with video and subtitles instead of MemN2N | MovieQA Video+Subtitles · MovieQA Video+Subtitles | accuracy: 34.2 -> 39.6 | Improved |
| Add 2-hop Static Word Memory (ablation) | LMN · MovieQA Video+Subtitles | accuracy: 39.3 -> 39.6; 39.6 -> 39.6 | No clear effect |
| Add 3-hop Static Word Memory (ablation) | LMN · MovieQA Video+Subtitles | accuracy: 39.3 -> 40.1; 39.6 -> 39.9 | Improved |
| Add 2 update-mechanism hops to Dynamic Subtitle Memory (ablation) | LMN · MovieQA Video+Subtitles | accuracy: 39.3 -> 41.4; 39.6 -> 41.6 | Improved |
| Use 3 update-mechanism hops in Dynamic Subtitle Memory instead of 2 (ablation) | LMN · MovieQA Video+Subtitles | accuracy: 41.4 -> 41.5; 41.6 -> 41.4 | No clear effect |
| Add question-guided subtitle attention (ablation) | LMN · MovieQA Video+Subtitles | accuracy: 39.3 -> 40.6; 39.6 -> 40.1 | Improved |
| Add question-guided subtitle attention to LMN with 2 update-mechanism hops (ablation) | LMN + UM² · MovieQA Video+Subtitles | accuracy: 41.4 -> 42.3; 41.6 -> 42.5 | Improved |
| Use LMN with update mechanism and question-guided attention instead of RWMN | MovieQA online evaluation · MovieQA online test Video+Subtitles | accuracy: 36.25 -> 39.03 | Improved |
Closest papers by meaning
- Progressive Attention Memory Network for Movie Story Question Answering (CVPR 2019 · similarity 0.91)
- A Read-Write Memory Network for Movie Story Understanding (ICCV 2017 · similarity 0.90)
- Knowledge-Based Video Question Answering with Unsupervised Scene Descriptions (ECCV 2020 · similarity 0.89)
- MovieChat+: Question-Aware Sparse Memory for Long Video Question Answering (TPAMI 2025 · similarity 0.88)
- LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding (NeurIPS 2024 · similarity 0.86)
- CALVIN: Improved Contextual Video Captioning via Instruction Tuning (NeurIPS 2024 · 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.
