Research map / Video / Video-language understanding
NExT-QA: Next Phase of Question-Answering to Explaining Temporal Actions
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 video features rather than no video clips (ablation) | HGA with BERT-FT · NExT-QA multi-choice QA | accuracy | Helped |
| Increase sampled video clips from 0 to 32 (ablation) | HGA with BERT-FT · NExT-QA multi-choice QA | accuracy: more than 15% | Helped |
| Replace C3D with I3D ResNeXt for motion features (ablation) | ResNet appearance plus motion features · NExT-QA multi-choice QA | accuracy | Helped |
| Fine-tune BERT for multi-choice QA instead of using off-the-shelf BERT (ablation) | BlindQA · NExT-QA validation | accuracy: 25.72 -> 43.76 | Helped |
| Use GloVe instead of off-the-shelf BERT representations (ablation) | BlindQA · NExT-QA validation | accuracy: 25.72 -> 30.60 | Helped |
| Add question attention to the answer decoder (ablation) | HGA open-ended QA · NExT-QA validation | WUPS: 20.44 -> 21.48 | Helped |
| Adapt strong multi-choice models to generate open-ended answers | VideoQA models · NExT-QA causal and temporal questions | WUPS | Hurt |
| Use heterogeneous graph reasoning for video-question relations | HGA · NExT-QA multi-choice QA | accuracy: 50.01 | Helped |
Closest papers by meaning
- MoReVQA: Exploring Modular Reasoning Models for Video Question Answering (CVPR 2024 · similarity 0.91)
- VideoEspresso: A Large-Scale Chain-of-Thought Dataset for Fine-Grained Video Reasoning via Core Frame Selection (CVPR 2025 · similarity 0.89)
- Can I Trust Your Answer? Visually Grounded Video Question Answering (CVPR 2024 · similarity 0.89)
- LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding (NeurIPS 2024 · similarity 0.88)
- Commonsense Video Question Answering through Video-Grounded Entailment Tree Reasoning (CVPR 2025 · similarity 0.88)
- Divide and Conquer: Exploring Language-centric Tree Reasoning for Video Question-Answering (ICML 2025 · similarity 0.88)
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.