Learning Stress Patterns with a Sequence-to-Sequence Neural Network
- Brandon Prickett (Linguistics Department, University of Massachusetts Amherst)
- Joe Pater
Abstract
We present the first application of modern neural networks to the well studied task of learning word stress systems. We tested our adaptation of a sequence-to-sequence network on the Tesar and Smolensky test set of 124 "languages", showing that it acquires generalizable representations of stress patterns in a very high proportion of runs. We also show that it learns restricted lexically conditioned patterns, known as stress windows. The ability of this model to acquire lexical idiosyncracies, which are very common in natural language systems, sets it apart from past, non-neural models tested on the Tesar and Smolensky data set.
Keywords: phonology, computational, stress, prosody, feet, hidden structure, neural networks
How to Cite:
Prickett, B. & Pater, J., (2022) “Learning Stress Patterns with a Sequence-to-Sequence Neural Network”, Society for Computation in Linguistics 5(1), 112-118. doi: https://doi.org/10.7275/xdbc-6925
Downloads:
Download PDF