Research map / Vision-language and multimodal / Visual question answering
Neuro-Symbolic Visual Reasoning: Disentangling "Visual" from "Reasoning"
Read the paperCheck 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 top-down contextual calibration to ∇-FOL (ablation) | ∇-FOL · GQA Test-Dev, All | accuracy: 51.45 % -> 54.76 % | Helped |
| Add top-down contextual calibration to ∇-FOL (ablation) | ∇-FOL · GQA Hard Test-Dev, All | accuracy: 1.81 % -> 12.91 % | Helped |
| Add top-down contextual calibration to ∇-FOL (ablation) | ∇-FOL · GQA Test-Dev, Binary | accuracy: 64.65 % -> 71.99 % | Helped |
| Add top-down contextual calibration to ∇-FOL (ablation) | ∇-FOL · GQA Hard Test-Dev, Binary | accuracy: 4.42 % -> 37.82 % | Helped |
| Add top-down contextual calibration to ∇-FOL (ablation) | ∇-FOL · GQA Test-Dev, Open | accuracy: 41.22 % -> 41.22 % | No clear effect |
| Add top-down contextual calibration to ∇-FOL (ablation) | ∇-FOL · GQA Hard Test-Dev, Open | accuracy: 0.53 % -> 0.53 % | No clear effect |
| Add top-down contextual calibration to ∇-FOL (ablation) | ∇-FOL · GQA Easy Test-Dev, All | error: 2.39 % -> 6.32 % | Hurt |
| Use LXMERT instead of MAC | VQA models using the same Faster-RCNN features · GQA Hard Test-Dev, All | accuracy: 30.54 % -> 38.43 % | Helped |
Closest papers by meaning
- Neural-Symbolic VQA: Disentangling Reasoning from Vision and Language Understanding (NeurIPS 2018 · similarity 0.94)
- Probabilistic Neural Symbolic Models for Interpretable Visual Question Answering (ICML 2019 · similarity 0.91)
- Learning to Reason Iteratively and Parallelly for Complex Visual Reasoning Scenarios (NeurIPS 2024 · similarity 0.90)
- Visual Graph Arena: Evaluating Visual Conceptualization of Vision and Multimodal Large Language Models (ICML 2025 · similarity 0.90)
- Visual Chain-of-Thought Prompting for Knowledge-Based Visual Reasoning (AAAI 2024 · similarity 0.89)
- When Open-Vocabulary Visual Question Answering Meets Causal Adapter: Benchmark and Approach (AAAI 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.