Abstract:An antithetical concept, adaptive symmetry, to conservative symmetry in physics is proposed to understand the deep neural networks (DNNs). It characterizes the invariance of variance, where a biotic system explores different pathways of evolution with equal probability in absence of feedback signals, and complex functional structure emerges from quantitative accumulation of adaptive-symmetries breaking in response to feedback signals. Theoretically and experimentally, we characterize the optimization process of a DNN system as an extended adaptive-symmetry-breaking process. One particular finding is that a hierarchically large DNN would have a large reservoir of adaptive symmetries, and when the information capacity of the reservoir exceeds the complexity of the dataset, the system could absorb all perturbations of the examples and self-organize into a functional structure of zero training errors measured by a certain surrogate risk. More specifically, this process is characterized by a statistical-mechanical model that could be appreciated as a generalization of statistics physics to the DNN organized complex system, and characterizes regularities in higher dimensionality. The model consists of three constitutes that could be appreciated as the counterparts of Boltzmann distribution, Ising model, and conservative symmetry, respectively: (1) a stochastic definition/interpretation of DNNs that is a multilayer probabilistic graphical model, (2) a formalism of circuits that perform biological computation, (3) a circuit symmetry from which self-similarity between the microscopic and the macroscopic adaptability manifests. The model is analyzed with a method referred as the statistical assembly method that analyzes the coarse-grained behaviors (over a symmetry group) of the heterogeneous hierarchical many-body interaction in DNNs.

Gamma-Hadron Discrimination with a Neural Network in the ARGO-YBJ Experiment

Replica symmetry breaking in supervised and unsupervised Hebbian networks

Correlations in Quantum Network Topologies Created with Cloning

Replica symmetry breaking in neural networks: a few steps toward rigorous results

Unsupervised and Supervised learning by Dense Associative Memory under replica symmetry breaking

Replica symmetric evaluation of the information transfer in a two-layer network in presence of continuous+discrete stimuli

Replica symmetric evaluation of the information transfer in a two-layer network in presence of continuous and discrete stimuli

Learning in the machine: The symmetries of the deep learning channel

Dense Hebbian neural networks: a replica symmetric picture of unsupervised learning

Complexity from Adaptive-Symmetries Breaking: Global Minima in the Statistical Mechanics of Deep Neural Networks

Correlations reveal the hierarchical organization of biological networks with latent variables

Symmetry Breaking and Equivariant Neural Networks

Echoes in correlated neural systems

Multilayer neural networks with extensively many hidden units

On the hardness of learning under symmetries

Correlative Information Maximization: A Biologically Plausible Approach to Supervised Deep Neural Networks without Weight Symmetry

Symmetry breaker governs synchrony patterns in neuronal inspired networks

Linking Network and Neuron-level Correlations by Renormalized Field Theory

Network Inference with Hidden Units

Solution space and storage capacity of fully connected two-layer neural networks with generic activation functions