Research map / Video / Video-language understanding
Uncovering the Temporal Context for Video Question Answering
Linchao Zhu, Zhongwen Xu, Yi Yang, Alexander G. Hauptmann
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 |
|---|---|---|---|
| replace CCA with dual-channel ranking objective | multiple-choice video QA · TACoS Present-Easy | accuracy: 65.1% -> 76.3% | Improved |
| replace CCA with dual-channel ranking objective | multiple-choice video QA · MPII-MD Present-Easy | accuracy: 41.6% -> 72.0% | Improved |
| replace CCA with dual-channel ranking objective | multiple-choice video QA · MEDTest 14 Present-Easy | accuracy: 63.2% -> 81.0% | Improved |
| replace GRU with LSTM in the encoder-decoder (ablation) | encoder-decoder video model · MEDTest 14 100Ex classification | accuracy: 20.4% -> 24.9% | Improved |
| replace averaged ConvNet features with GRU encoder-decoder features | dual-channel ranking · TACoS Past-Easy | accuracy: 74.8% -> 78.3% | Improved |
| replace averaged ConvNet features with GRU encoder-decoder features | dual-channel ranking · TACoS Present-Easy | accuracy: 76.3% -> 79.7% | Improved |
| replace averaged ConvNet features with GRU encoder-decoder features | dual-channel ranking · TACoS Future-Easy | accuracy: 76.4% -> 78.7% | Improved |
| replace averaged ConvNet features with GRU encoder-decoder features | dual-channel ranking · MPII-MD Past-Easy | accuracy: 66.8% -> 72.1% | Improved |
| replace averaged ConvNet features with GRU encoder-decoder features | dual-channel ranking · MPII-MD Future-Easy | accuracy: 68.7% -> 73.6% | Improved |
| replace averaged ConvNet features with GRU encoder-decoder features | dual-channel ranking · MPII-MD Present-Hard | accuracy: 47.3% -> 48.2% | Improved |
| add the word channel to sentence-only ranking (vary lambda) (ablation) | dual-channel ranking · Present-Easy | accuracy | Improved |
| predict future clip instead of inferring past clip | GRU encoder-decoder · TACoS and MPII-MD | accuracy: Past-Easy 78.1% -> Future-Easy 76.9% | Mixed |
Closest papers by meaning
- Temporal Chain of Thought: Long-Video Understanding by Thinking in Frames (NeurIPS 2025 · similarity 0.91)
- Language-aware Visual Semantic Distillation for Video Question Answering (CVPR 2024 · similarity 0.90)
- Not Another Text Benchmark: Putting the “Visual" Back in Visual Question Answering for Large Video Models (NeurIPS 2026 · similarity 0.90)
- LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding (NeurIPS 2024 · similarity 0.90)
- ReWind: Understanding Long Videos with Instructed Learnable Memory (CVPR 2025 · similarity 0.89)
- Streaming Long Video Understanding with Large Language Models (NeurIPS 2024 · similarity 0.89)
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.
