A Feedforward Unitary Equivariant Neural Network

A Feedforward Unitary Equivariant Neural Network

A Feedforward Unitary Equivariant Neural Network 150 150 UKAEA Opendata
UKAEA-CCFE-PR(22)62

A Feedforward Unitary Equivariant Neural Network

We have devised a new type of feedforward neural network. It is equivariant with respect to unitary operators $U(n)$. The input and output can be vectors in $\\mathbb{C}^n$ with arbitrary dimension~$n$. No convolution layer is required in our implementation, and we also avoid errors due to truncated higher order terms in Fourier-like transformation. The implementation of each layer can be done efficiently using simple calculations. As a proof of concept, we have given empirical results on the prediction of the dynamics of atomic motion to demonstrate the practicality of our approach.

Collection:
Journals
Journal:
Neural Networks
Publisher:
Elsevier
Published date:
04/02/2023