Research map / Probabilistic and scientific modeling / Molecules and proteins
Disentangled Counterfactual Learning for Physical Audiovisual Commonsense Reasoning
Changsheng Lv, Shuai Zhang, Yapeng Tian, Mengshi Qi, Huadong Ma
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 DCL | Late Fusion · PACS | accuracy: 55.0 -> 57.7 | Improved |
| Add DCL | CLIP · PACS | accuracy: 56.3 -> 58.4 | Improved |
| Add DCL | AudioCLIP · PACS | accuracy: 60.0 -> 63.2 | Improved |
| Add DCL | UNITER · PACS | accuracy: 60.6 -> 62.0 | Improved |
| Use AudioCLIP (with audio) instead of CLIP, both with DCL | DCL · PACS | accuracy: 58.4 -> 63.2 | Improved |
| Add DSE (ablation) | AudioCLIP · PACS | accuracy: 60.0 -> 61.1 | Improved |
| Add DSE (ablation) | Late Fusion · PACS | accuracy: 55.0 -> 56.2 | Improved |
| Add MLP with same parameters as DSE instead (ablation) | Late Fusion · PACS | accuracy: 55.0 -> 54.9 | No clear effect |
| Add MLP instead of DSE (ablation) | AudioCLIP · PACS | accuracy: 60.0 -> 60.3 | No clear effect |
| Add physical knowledge relationship A on top of DSE (ablation) | CLIP · PACS | accuracy: 57.0 -> 57.8 | Improved |
| Add counterfactual relation intervention C on top of DSE,A (ablation) | AudioCLIP · PACS | accuracy: 61.9 -> 63.2 | Improved |
| Add C on top of DSE,A (ablation) | AudioCLIP · PACS-Material | accuracy: 75.8 -> 76.2 | Mixed |
Closest papers by meaning
- Counterfactual Debiasing for Physical Audiovisual Commonsense Reasoning (AAAI 2025 · similarity 0.92)
- What If the TV Was Off? Examining Counterfactual Reasoning Abilities of Multi-modal Language Models (CVPR 2024 · similarity 0.85)
- Evidence-RL: Towards Evidence-intensive Visual Reasoning (NeurIPS 2026 · similarity 0.84)
- When Eyes and Ears Disagree: Can MLLMs Discern Audio-Visual Confusion? (AAAI 2026 · similarity 0.84)
- CF-VLM:CounterFactual Vision-Language Fine-tuning (NeurIPS 2025 · similarity 0.84)
- Counterfactual reasoning: an analysis of in-context emergence (NeurIPS 2025 · similarity 0.84)
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.
