Research map / Vision-language and multimodal / Visual question answering
COCA: COllaborative CAusal Regularization for Audio-Visual Question Answering
Mingrui Lao, Nan Pu, Yu Liu, Kai He, Erwin M. Bakker, Michael S. Lew
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 COCA to AVST | AVST on MUSIC-AVQA · MUSIC-AVQA test | average accuracy: 71.52 -> 72.33 | Improved |
| Add COCA to the baseline | Baseline on MUSIC-AVQA · MUSIC-AVQA test | average accuracy: 69.61 -> 71.64 | Improved |
| Add COCA regularization | VQAT-based model on MUSIC-AVQA · MUSIC-AVQA test | average accuracy: 69.32 -> 70.81 | Improved |
| Add COCA regularization | ConvLSTM-based model on MUSIC-AVQA · MUSIC-AVQA test | average accuracy: 68.37 -> 69.74 | Improved |
| Add counterfactual regularization for the Q→G language shortcut (ablation) | Baseline AVQA model · MUSIC-AVQA test | average accuracy: 69.61 -> 70.50 | Improved |
| Add factual regularization for Q→G (ablation) | Baseline AVQA model with Q→G counterfactual regularization · MUSIC-AVQA test | average accuracy: 70.50 -> 70.86 | Improved |
| Add counterfactual regularization for the F→G visual shortcut (ablation) | Baseline AVQA model · MUSIC-AVQA test | average accuracy: 69.61 -> 69.91 | Improved |
| Add factual regularization for F→G (ablation) | Baseline AVQA model with F→G counterfactual regularization · MUSIC-AVQA test | average accuracy: 69.91 -> 70.12 | Improved |
| Add counterfactual regularization for the S→G audio shortcut (ablation) | Baseline AVQA model · MUSIC-AVQA test | average accuracy: 69.61 -> 70.04 | Improved |
| Add factual regularization for S→G (ablation) | Baseline AVQA model with S→G counterfactual regularization · MUSIC-AVQA test | average accuracy: 70.04 -> 70.37 | Improved |
| Add counterfactual regularization for all three shortcut biases (ablation) | Baseline AVQA model · MUSIC-AVQA test | average accuracy: 69.61 -> 70.81 | Improved |
| Add factual regularization for all three shortcut biases (ablation) | Multiple-shortcut counterfactual model · MUSIC-AVQA test | average accuracy: 70.81 -> 71.21 | Improved |
Later papers that compared with this method
21 later accepted papers report CoCA as a baseline in their results tables; 15 of them report a better number on most of the results they share with it.
- Attention Misses Visual Risk: Risk-Adaptive Steering for Multimodal Safety Alignment (ECCV 2026 · better on 1 of 1 shared result)
- SafePTR: Token-Level Jailbreak Defense in Multimodal LLMs via Prune-then-Restore Mechanism (NeurIPS 2025 · better on 2 of 2 shared results)
- Insect-Foundation: A Foundation Model and Large Multimodal Dataset for Vision-Language Insect Understanding (IJCV 2025 · better on 2 of 2 shared results)
- BiMAC: Bidirectional Multimodal Alignment in Contrastive Learning (AAAI 2025 · better on 1 of 1 shared result)
- Immune: Improving Safety Against Jailbreaks in Multi-modal LLMs via Inference-Time Alignment (CVPR 2025 · better on 1 of 1 shared result)
- Learning Visual Composition through Improved Semantic Guidance (CVPR 2025 · better on 0 of 1 shared result)
- Derm1M: A Million-scale Vision-Language Dataset Aligned with Clinical Ontology Knowledge for Dermatology (ICCV 2025 · better on 1 of 1 shared result)
- SyCoCa: Symmetrizing Contrastive Captioners with Attentive Masking for Multimodal Alignment (ICML 2024 · better on 12 of 12 shared results)
13 more not listed.
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Look, Listen, and Answer: Overcoming Biases for Audio-Visual Question Answering (NeurIPS 2024 · similarity 0.93)
- Audio-Visual Adaptive Fusion Network for Question Answering Based on Contrastive Learning (AAAI 2025 · similarity 0.89)
- When Open-Vocabulary Visual Question Answering Meets Causal Adapter: Benchmark and Approach (AAAI 2025 · similarity 0.89)
- JointAVBench: A Benchmark for Joint Audio-Visual Reasoning Evaluation (ICLR 2026 · similarity 0.89)
- Question-Aware Gaussian Experts for Audio-Visual Question Answering (CVPR 2025 · similarity 0.88)
- AVQACL: A Novel Benchmark for Audio-Visual Question Answering Continual Learning (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.
