Emergent field theories from neural networks
This work provides a theoretical framework linking physics and machine learning, potentially enabling new modeling approaches, but it appears incremental as it builds on known dualities.
The authors established a duality between Hamiltonian systems and neural networks, showing that Hamilton's equations correspond to activation and learning dynamics, and applied this to model field theories like Klein-Gordon and Dirac fields with specific tensor properties.
We establish a duality relation between Hamiltonian systems and neural network-based learning systems. We show that the Hamilton's equations for position and momentum variables correspond to the equations governing the activation dynamics of non-trainable variables and the learning dynamics of trainable variables. The duality is then applied to model various field theories using the activation and learning dynamics of neural networks. For Klein-Gordon fields, the corresponding weight tensor is symmetric, while for Dirac fields, the weight tensor must contain an anti-symmetric tensor factor. The dynamical components of the weight and bias tensors correspond, respectively, to the temporal and spatial components of the gauge field.