Deep Learning of Compositional Targets with Hierarchical Spectral Methods
stat.ML, cs.LG
Submitted: 2026-02-11
Updated: 2026-11-03
Terminology
Sources
- Repetita Iuvant: Data Repetition Allows SGD to Learn High-Dimensional Multi-Index Functions
- Optimal Spectral Transitions in High-Dimensional Multi-Index Models
- A Precise Performance Analysis of Learning with Random Features
- Asymptotics of Random Feature Regression Beyond the Linear Scaling Regime
- Adam: A Method for Stochastic Optimization
- Spectral Estimators for Multi-Index Models: Precise Asymptotics and Optimal Weak Recovery
- An exact mapping between the Variational Renormalization Group and Deep Learning
- Deep Learning and Hierarchal Generative Models
- Fundamental computational limits of weak learnability in high-dimensional multi-index models
- Introduction to the non-asymptotic analysis of random matrices
- Learning Hierarchical Polynomials with Three-Layer Neural Networks
- When does Gaussian equivalence fail and how to fix it: Non-universal behavior of random features with quadratic scaling
- Neural Networks Learn Generic Multi-Index Models Near Information-Theoretic Limit
Related papers
- Uniform-in-Time Weak Propagation-of-Chaos in Shallow Neural Networks
- Diffusion Models Observe Only Gradients: A Geometric Perspective on Score Matching Errors
- Quantifying Error Propagation and Model Collapse in Diffusion Models
- Behavior of prediction performance metrics with rare events
- Statistical Inference and Learning for Shapley Additive Explanations (SHAP)
- Control Variate Design for Score-Function Gradient Estimators in Importance-Weighted Variational Inference