arXiv · 2305.02334
Structures of Neural Network Effective Theories
Abstract
We develop a diagrammatic approach to effective field theories (EFTs) corresponding to deep neural networks at initialization, which dramatically simplifies computations of finite-width corrections to neuron statistics. The structures of EFT calculations make it transparent that a single condition governs criticality of all connected correlators of neuron preactivations. Understanding of such EFTs may facilitate progress in both deep learning and field theory simulations.
Explore related subjects
Keep this discovery
Ian Banta, Tianji Cai, Nathaniel Craig, Zhengkang Zhang. 2023-05-03. Structures of Neural Network Effective Theories. https://arxiv.org/abs/2305.02334
Cite the original work for its findings. Save a collection to share your selection of sources.