Research map / Learning theory and algorithms / Kernels and regression theory
A Spectral Theory of Neural Prediction and Alignment
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 |
|---|---|---|---|
| Use trained rather than randomly initialized network representations | DNN encoding models · V2, V4, and IT neural prediction | neural prediction error | Helped |
| Use trained rather than randomly initialized network representations | DNN encoding models · V1 neural prediction | neural prediction error | No clear effect |
| Use trained-model alignment coefficients with either trained or random ResNet50 eigenspectra (ablation) | ResNet50 · IT neural prediction | error mode radius | Helped |
| Use early-stage rather than late-stage eigenspectra (ablation) | DNN encoding models · V1 neural prediction | neural prediction error | Helped |
| Use late-stage rather than early-stage alignment coefficients (ablation) | DNN encoding models · IT neural prediction | neural prediction error | Helped |
| Use early rather than late model stages | DNN encoding models · V1 neural prediction | neural prediction error | Helped |
| Use late rather than early model stages | DNN encoding models · IT neural prediction | neural prediction error | Helped |
| Use robust rather than standard training | ResNet50 · V1 neural prediction, early stages | neural prediction error | Helped |
| Use robust rather than standard training | ResNet50 · V1 neural prediction, late stages | neural prediction error | No clear effect |
| Use robust rather than standard training | ResNet50 · V2, V4, and IT neural prediction | neural prediction error | No clear effect |
| Use robust-model alignment coefficients rather than standard-model alignment coefficients (ablation) | ResNet50 · V1 neural prediction | neural prediction error | Helped |
| Use robust-model eigenspectra rather than standard-model eigenspectra (ablation) | ResNet50 · V1 neural prediction, early stages | neural prediction error | No clear effect |
Closest papers by meaning
- Spectral Analysis of Representational Similarity with Limited Neurons (NeurIPS 2025 · similarity 0.88)
- Bridging Critical Gaps in Convergent Learning: How Representational Alignment Evolves Across Layers, Training, and Distribution Shifts (NeurIPS 2025 · similarity 0.86)
- Model–Behavior Alignment under Flexible Evaluation: When the Best-Fitting Model Isn’t the Right One (NeurIPS 2025 · similarity 0.86)
- Understanding Sample Efficiency in Predictive Coding (NeurIPS 2026 · similarity 0.86)
- Not all solutions are created equal: An analytical dissociation of functional and representational similarity in deep linear neural networks (ICML 2025 · similarity 0.86)
- Analyzing Neural Scaling Laws in Two-Layer Networks with Power-Law Data Spectra (ICLR 2025 · similarity 0.86)
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.