Steps Toward Deep Kernel Methods from Infinite Neural Networks
This work provides a theoretical explanation for the empirical success of deep neural networks, which is foundational for the ML/AI community.
The paper tackled the generalization behavior of deep neural networks by analyzing deep infinite layers, showing they align with Gaussian processes and kernel methods, and devised stochastic kernels that encode network information while maintaining stability.
Contemporary deep neural networks exhibit impressive results on practical problems. These networks generalize well although their inherent capacity may extend significantly beyond the number of training examples. We analyze this behavior in the context of deep, infinite neural networks. We show that deep infinite layers are naturally aligned with Gaussian processes and kernel methods, and devise stochastic kernels that encode the information of these networks. We show that stability results apply despite the size, offering an explanation for their empirical success.