Dataflow Matrix Machines as a Generalization of Recurrent Neural Networks
Neural and Evolutionary Computing
2018-05-29 v2
Abstract
Dataflow matrix machines are a powerful generalization of recurrent neural networks. They work with multiple types of arbitrary linear streams, multiple types of powerful neurons, and allow to incorporate higher-order constructions. We expect them to be useful in machine learning and probabilistic programming, and in the synthesis of dynamic systems and of deterministic and probabilistic programs.
Keywords
Cite
@article{arxiv.1603.09002,
title = {Dataflow Matrix Machines as a Generalization of Recurrent Neural Networks},
author = {Michael Bukatin and Steve Matthews and Andrey Radul},
journal= {arXiv preprint arXiv:1603.09002},
year = {2018}
}
Comments
4 pages position paper (v2 - update references)