Research map / Vision-language and multimodal / Human and embodied modeling
Uncovering Semantic Selectivity of Latent Groups in Higher Visual Cortex with Mutual Information-Guided Diffusion
Yule Wang, Joseph Yu, Chengrui Li, Weihan Li, Anqi 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 |
|---|---|---|---|
| add weak label supervision and partial correlation penalty to the VAE | neural VAE · IT cortex reconstruction (M1) | R2 (%): 78.62 -> 76.58 | Hurt |
| add weak label supervision and partial correlation penalty to the VAE | neural VAE · IT cortex reconstruction (M2) | R2 (%): 83.72 -> 81.86 | Hurt |
| add weak label supervision and partial correlation penalty to the VAE | neural VAE · IT cortex disentanglement (M1) | MIG (%): 33.27 -> 44.23 | Improved |
| add weak label supervision and partial correlation penalty to the VAE | neural VAE · IT cortex disentanglement (M2) | MIG (%): 28.65 -> 49.85 | Improved |
| add weak label supervision to the VAE (ablation) | MIG-Vis VAE · IT cortex reconstruction (M1) | R2 (%): 76.90 -> 76.58 | Hurt |
| add weak label supervision to the VAE (ablation) | MIG-Vis VAE · IT cortex reconstruction (M2) | R2 (%): 82.39 -> 81.86 | Hurt |
| add partial correlation penalty to the VAE (ablation) | MIG-Vis VAE · IT cortex reconstruction (M1) | R2 (%): 77.30 -> 76.58 | Hurt |
| add partial correlation penalty to the VAE (ablation) | MIG-Vis VAE · IT cortex reconstruction (M2) | R2 (%): 82.21 -> 81.86 | Hurt |
| add weak label supervision and partial correlation penalty to the VAE | neural VAE · IT cortex reconstruction (M1) | RMSE: 48.49 -> 50.47 | Hurt |
| add weak label supervision and partial correlation penalty to the VAE | neural VAE · IT cortex reconstruction (M2) | RMSE: 40.77 -> 43.44 | Hurt |
| add mutual information guidance instead of likelihood-based guidance in diffusion (ablation) | MIG-Vis · Latent group 2 inter-category synthesis (qualitative) | visual quality of category transitions | Improved |
| add mutual information guidance instead of likelihood-based guidance in diffusion (ablation) | MIG-Vis · Latent group 1 pose synthesis (qualitative) | visual rotation quality | No clear effect |
Closest papers by meaning
- In Silico Mapping of Visual Categorical Selectivity Across the Whole Brain (NeurIPS 2025 · similarity 0.90)
- BrainSCUBA: Fine-Grained Natural Language Captions of Visual Cortex Selectivity (ICLR 2024 · similarity 0.89)
- Finding Shared Decodable Concepts and their Negations in the Brain (ICLR 2025 · similarity 0.89)
- Brain Mapping with Dense Features: Grounding Cortical Semantic Selectivity in Natural Images With Vision Transformers (ICLR 2025 · similarity 0.88)
- Sparse components distinguish visual pathways & their alignment to neural networks (ICLR 2025 · similarity 0.88)
- Flexible Context-Driven Sensory Processing in Dynamical Vision Models (NeurIPS 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.
