Research map / Vision-language and multimodal / Visual question answering
Compositional Neuro-Symbolic Concepts in Neural Activities
Yanchen Wang, Joy Hsu, Ehsan Adeli, Jiajun Wu
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 |
|---|---|---|---|
| use neuro-symbolic compositional grounding instead of the best prior decoder | fMRI decoder · BOLD5000-QA | accuracy (overall): 0.4773 -> 0.7041 | Improved |
| use neuro-symbolic compositional grounding instead of the best prior decoder | fMRI decoder · CNeuroMod-QA | accuracy (overall): 0.4642 -> 0.7046 | Improved |
| use NEURONA on unseen query compositions | fMRI decoder · BOLD5000-QA (unseen queries) | accuracy (overall): 0.4754 -> 0.6840 | Improved |
| use NEURONA on unseen query compositions | fMRI decoder · CNeuroMod-QA (unseen queries) | accuracy (overall): 0.4502 -> 0.6583 | Improved |
| add multi-region grounding without guidance (ablation) | NEURONA single-region · BOLD5000-QA | accuracy (overall): 0.6451 -> 0.6476 | No clear effect |
| add multi-region grounding without guidance (ablation) | NEURONA single-region · CNeuroMod-QA | accuracy (overall): 0.6429 -> 0.6162 | Hurt |
| add subject guidance to multi-region predicate grounding (ablation) | NEURONA multi-region · BOLD5000-QA | accuracy (overall): 0.6476 -> 0.6678 | Improved |
| add object guidance to multi-region predicate grounding (ablation) | NEURONA multi-region · CNeuroMod-QA | accuracy (overall): 0.6162 -> 0.6872 | Improved |
| add both subject and object guidance (ablation) | NEURONA multi-region · BOLD5000-QA | accuracy (overall): 0.6476 -> 0.7102 | Improved |
| add both subject and object guidance (ablation) | NEURONA multi-region · CNeuroMod-QA | accuracy (overall): 0.6162 -> 0.7189 | Improved |
| use Yeo-17 atlas instead of Yeo-7 (ablation) | NEURONA · BOLD5000-QA | accuracy (overall): 0.6864 -> 0.7041 | Improved |
| use all Yeo-7 networks instead of one network subset (ablation) | NEURONA · BOLD5000-QA | accuracy (overall): 0.6804 -> 0.7102 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- UMBRAE — UMBRAE: Unified Multimodal Brain Decoding (ECCV 2024) · better on 5 of 5 shared results
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
- NeSyCoCo: A Neuro-Symbolic Concept Composer for Compositional Generalization (AAAI 2025 · similarity 0.88)
- Compositional Explanations of Neurons (NeurIPS 2020 · similarity 0.87)
- Far from the Shallow: Brain-Predictive Reasoning Embedding through Residual Disentanglement (NeurIPS 2025 · similarity 0.86)
- MindLLM: A Subject-Agnostic and Versatile Model for fMRI-to-text Decoding (ICML 2025 · similarity 0.85)
- Neuro-Vision to Language: Enhancing Brain Recording-based Visual Reconstruction and Language Interaction (NeurIPS 2024 · similarity 0.85)
- Concise and Logically Consistent Conformal Sets for Neuro-Symbolic Concept-Based Models (NeurIPS 2026 · similarity 0.85)
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.
