Dataflow Matrix Machines as a Generalization of Recurrent Neural Networks
Michael A. Bukatin, Matthews, Steve, Andrey Radul · arXiv (Cornell University) · 2016
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.