Echo State Neural Machine Translation
This work demonstrates that randomized networks can perform well on complex NLP tasks, though it is incremental in applying existing echo state network concepts to machine translation.
The authors tackled neural machine translation by using echo state networks with fixed random weights, achieving 70-80% of the quality of fully trainable baselines.
We present neural machine translation (NMT) models inspired by echo state network (ESN), named Echo State NMT (ESNMT), in which the encoder and decoder layer weights are randomly generated then fixed throughout training. We show that even with this extremely simple model construction and training procedure, ESNMT can already reach 70-80% quality of fully trainable baselines. We examine how spectral radius of the reservoir, a key quantity that characterizes the model, determines the model behavior. Our findings indicate that randomized networks can work well even for complicated sequence-to-sequence prediction NLP tasks.