Research map / Label-efficient and robust learning / Recognition and long-tail
Brain-Mediated Transfer Learning of Convolutional Neural Networks
Satoshi Nishida, Yusuke Nakano, Antoine Blanc, Naoya Maeda, Masataka Kado, Shinji Nishimoto
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 |
|---|---|---|---|
| transform VGG-16 features into brain-response representations with BTL instead of using direct TL | Scene description estimation · Scene descriptions (Task1) | estimation performance: 0.537 -> 0.546 | Improved |
| transform CNN features into brain-response representations with BTL instead of direct TL | Impression rating estimation · Impression ratings (Task2) | estimation performance: 0.501 -> 0.517 | Improved |
| use BTL rather than direct TL | Click rate estimation · Click rate (Task3-1) | estimation performance: 0.510 -> 0.505 | Mixed |
| use brain-mediated features rather than direct CNN features | View completion rate estimation · View completion rate (Task3-2) | estimation performance: 0.303 -> 0.444 | Improved |
| use BTL rather than direct TL | Ad preference estimation · Ad preference (Task4) | estimation performance: 0.365 -> 0.387 | Improved |
| add SoundNet acoustic features (ablation) | BTL impression rating estimation · Impression ratings (Task2) | estimation performance: 0.500 -> 0.517 | Improved |
| add SoundNet acoustic features (ablation) | BTL scene description estimation · Scene descriptions (Task1) | estimation performance: 0.546 -> 0.536 | Hurt |
| add vox2vox (ablation) | BTL voxel-response prediction with VGG-16 and SoundNet · Web ad movies test fMRI | voxel-response prediction accuracy: 0.106 -> 0.108 | Improved |
| add vox2vox (ablation) | BTL impression rating estimation · Impression ratings (Task2) | estimation performance: 0.516 -> 0.517 | Improved |
| add vox2vox (ablation) | BTL click rate estimation with VGG-16 · Click rate (Task3-1) | estimation performance: 0.505 -> 0.502 | Hurt |
| add vox2vox (ablation) | BTL view completion estimation with VGG-16 · View completion rate (Task3-2) | estimation performance: 0.440 -> 0.444 | Improved |
| add vox2vox (ablation) | BTL ad preference estimation with VGG-16 and SoundNet · Ad preference (Task4) | estimation performance: 0.381 -> 0.387 | Improved |
Closest papers by meaning
- Toward Generalizing Visual Brain Decoding to Unseen Subjects (ICLR 2025 · similarity 0.85)
- Meta-Learning an In-Context Transformer Model of Human Higher Visual Cortex (NeurIPS 2025 · similarity 0.85)
- Brain-tuning Improves Generalizability and Efficiency of Brain Alignment in Speech Models (NeurIPS 2025 · similarity 0.84)
- Understanding the Transferability of Representations via Task-Relatedness (NeurIPS 2024 · similarity 0.84)
- Evaluating alignment between humans and neural network representations in image-based learning tasks (NeurIPS 2024 · similarity 0.84)
- See Through Their Minds: Learning Transferable Brain Decoding Models from Cross-Subject fMRI (AAAI 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.
