Research map / Vision-language and multimodal / Visual question answering
CRIC: A VQA Dataset for Compositional Reasoning on Vision and Commonsense
Difei Gao, Ruiping Wang, Shiguang Shan, Xilin Chen
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 attention supervision loss (ablation) | ViLBERT · CRIC test | grounding accuracy (%): 27.06 -> 62.79 | Improved |
| add attention supervision loss (ablation) | ViLBERT · CRIC test | final score (%): 23.67 -> 53.76 | Improved |
| add attention supervision loss (ablation) | ViLBERT · CRIC test | answer accuracy (%): 76.07 -> 77.54 | Improved |
| add attention supervision loss (ablation) | Memory-VQA · CRIC test | grounding accuracy (%): 27.67 -> 49.51 | Improved |
| add attention supervision loss (ablation) | Memory-VQA · CRIC test | answer accuracy (%): 66.59 -> 66.93 | Improved |
| add attention supervision loss (ablation) | Bottom-Up · CRIC test | grounding accuracy (%): 19.88 -> 38.10 | Improved |
| add attention supervision loss (ablation) | Bottom-Up · CRIC test | answer accuracy (%): 64.71 -> 62.39 | Hurt |
| add explicit knowledge-item branch (Memory-VQA vs Bottom-Up) | Bottom-Up · CRIC test | grounding accuracy (%): 19.88 -> 27.67 | Improved |
| add explicit knowledge-item branch (Memory-VQA vs Bottom-Up) | Bottom-Up · CRIC test | answer accuracy (%): 64.71 -> 66.59 | Improved |
| use compositional modular model (MAC) instead of monolithic Bottom-Up | Bottom-Up · CRIC test | answer accuracy (%): 64.71 -> 68.91 | Improved |
| use pretrained ViLBERT instead of Memory-VQA | Memory-VQA · CRIC test | answer accuracy (%): 66.59 -> 76.07 | Improved |
Closest papers by meaning
- Compositional Substitutivity of Visual Reasoning for Visual Question Answering (ECCV 2024 · similarity 0.91)
- GQA: A New Dataset for Real-World Visual Reasoning and Compositional Question Answering (CVPR 2019 · similarity 0.91)
- ConMe: Rethinking Evaluation of Compositional Reasoning for Modern VLMs (NeurIPS 2024 · similarity 0.90)
- Natural Language Inference Improves Compositionality in Vision-Language Models (ICLR 2025 · similarity 0.89)
- Core-to-Global Reasoning for Compositional Visual Question Answering (AAAI 2025 · similarity 0.89)
- Investigating Compositional Challenges in Vision-Language Models for Visual Grounding (CVPR 2024 · 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.
