Mehler's Formula, Branching Process, and Compositional Kernels of Deep Neural Networks
arXiv:2004.04767 · doi:10.1080/01621459.2020.1853547
Abstract
We utilize a connection between compositional kernels and branching processes via Mehler's formula to study deep neural networks. This new probabilistic insight provides us a novel perspective on the mathematical role of activation functions in compositional neural networks. We study the unscaled and rescaled limits of the compositional kernels and explore the different phases of the limiting behavior, as the compositional depth increases. We investigate the memorization capacity of the compositional kernels and neural networks by characterizing the interplay among compositional depth, sample size, dimensionality, and non-linearity of the activation. Explicit formulas on the eigenvalues of the compositional kernel are provided, which quantify the complexity of the corresponding reproducing kernel Hilbert space. On the methodological front, we propose a new random features algorithm, which compresses the compositional layers by devising a new activation function.
References in corpus (11)
- Understanding deep learning requires rethinking generalization
- A Mean Field View of the Landscape of Two-Layers Neural Networks
- Gradient Descent Provably Optimizes Over-parameterized Neural Networks
- The generalization error of random features regression: Precise asymptotics and double descent curve
- Scaling Limits of Wide Neural Networks with Weight Sharing: Gaussian Process Behavior, Gradient Independence, and Neural Tangent Kernel Derivation
- To understand deep learning we need to understand kernel learning
- Just Interpolate: Kernel "Ridgeless" Regression Can Generalize
- Does data interpolation contradict statistical optimality?
- Kernel and Rich Regimes in Overparametrized Models
- Optimal Regularization Can Mitigate Double Descent
- Training Neural Networks as Learning Data-adaptive Kernels: Provable Representation and Approximation Benefits