Research map / Video / Video-language understanding
AGQA: A Benchmark for Compositional Spatio-Temporal Reasoning
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 |
|---|---|---|---|
| Add visual features to HCRN (ablation) | HCRN · AGQA balanced, all questions | accuracy: 47.00 -> 47.42 | Helped |
| Train and test on held-out novel concept compositions | PSAC, HME, and HCRN · AGQA novel compositions | accuracy: HME: 47.74 -> 40.11 | Hurt |
| Train on simpler questions and test on questions with more reasoning steps | PSAC, HME, and HCRN · AGQA more compositional steps | accuracy: HME: 47.74 -> 39.70 | Hurt |
| Use visual features rather than linguistic features alone (ablation) | HCRN · AGQA balanced, all questions | accuracy: 0.42% | Helped |
Closest papers by meaning
- Tool-Augmented Spatiotemporal Reasoning for Streamlining Video Question Answering Task (NeurIPS 2025 · similarity 0.91)
- STEP: Enhancing Video-LLMs' Compositional Reasoning by Spatio-Temporal Graph-guided Self-Training (CVPR 2025 · similarity 0.91)
- SOK-Bench: A Situated Video Reasoning Benchmark with Aligned Open-World Knowledge (CVPR 2024 · similarity 0.90)
- SpatiaLQA: A Benchmark for Evaluating Spatial Logical Reasoning in Vision-Language Models (CVPR 2026 · similarity 0.90)
- VideoEspresso: A Large-Scale Chain-of-Thought Dataset for Fine-Grained Video Reasoning via Core Frame Selection (CVPR 2025 · similarity 0.89)
- VELOCITI: Benchmarking Video-Language Compositional Reasoning with Strict Entailment (CVPR 2025 · 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.