Research map / Vision-language and multimodal / Visual question answering
Q-TriM: Question-Guided Tri-Modal Attention for Audio–Visual Question Answering
SungHun Kim, Seung Baek
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 token filtering (ablation) | Q-TriM · MUSIC-AVQA-R | avg accuracy (%): 69.21 -> 70.89 | Improved |
| add Tri-Modal attention (ablation) | Q-TriM · MUSIC-AVQA-R | avg accuracy (%): 68.92 -> 70.89 | Improved |
| add modality mixture (ablation) | Q-TriM · MUSIC-AVQA-R | avg accuracy (%): 67.85 -> 70.89 | Improved |
| add all three modules to the base (ablation) | Q-TriM · MUSIC-AVQA-R | avg accuracy (%): 66.96 -> 70.89 | Improved |
| add token filtering (ablation) | Q-TriM · MUSIC-AVQA | avg accuracy (%): 75.58 -> 77.68 | Improved |
| add Tri-Modal attention (ablation) | Q-TriM · MUSIC-AVQA | avg accuracy (%): 75.41 -> 77.68 | Improved |
| use Tri-Modal attention instead of plain cross attention (ablation) | Q-TriM · MUSIC-AVQA | avg accuracy (%): 76.50 -> 77.68 | Improved |
| add modality mixture (ablation) | Q-TriM · MUSIC-AVQA | avg accuracy (%): 75.01 -> 77.68 | Improved |
| keep only 10% of patches instead of 50% (ablation) | Q-TriM token filtering · MUSIC-AVQA-R | avg accuracy (%): 70.89 -> 70.11 | Mixed |
| vary token keep ratio 0.1 to 0.9 (ablation) | Q-TriM token filtering · MUSIC-AVQA-R | avg accuracy (%): 70.11 / 70.56 / 70.89 / 70.39 / 70.50 | No clear effect |
| use adaptive keep ratio instead of fixed (ablation) | Q-TriM token filtering · MUSIC-AVQA | accuracy (%): 77.68 -> 76.02 | Hurt |
| use adaptive keep ratio instead of fixed (ablation) | Q-TriM token filtering · MUSIC-AVQA-R | accuracy (%): 70.89 -> 70.57 | Hurt |
Closest papers by meaning
- Audio-Visual Adaptive Fusion Network for Question Answering Based on Contrastive Learning (AAAI 2025 · similarity 0.91)
- Look, Listen, and Answer: Overcoming Biases for Audio-Visual Question Answering (NeurIPS 2024 · similarity 0.91)
- AVQACL: A Novel Benchmark for Audio-Visual Question Answering Continual Learning (CVPR 2025 · similarity 0.90)
- Question-Aware Gaussian Experts for Audio-Visual Question Answering (CVPR 2025 · similarity 0.89)
- Object-Aware Adaptive-Positivity Learning for Audio-Visual Question Answering (AAAI 2024 · similarity 0.89)
- FRAMES-VQA: Benchmarking Fine-Tuning Robustness across Multi-Modal Shifts in Visual Question Answering (CVPR 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.
