Inherent Biases of Recurrent Neural Networks for Phonological Assimilation and Dissimilation
This work addresses how neural networks can mimic human learning biases in phonology, but it is incremental as it builds on existing models by removing the need for additional feature representations.
The authors tackled the problem of modeling phonological pattern learning biases in neural networks, showing that a simple recurrent neural network learns single-feature patterns faster than two-feature patterns and vowel/consonant-only patterns faster than mixed ones, matching human experimental results.
A recurrent neural network model of phonological pattern learning is proposed. The model is a relatively simple neural network with one recurrent layer, and displays biases in learning that mimic observed biases in human learning. Single-feature patterns are learned faster than two-feature patterns, and vowel or consonant-only patterns are learned faster than patterns involving vowels and consonants, mimicking the results of laboratory learning experiments. In non-recurrent models, capturing these biases requires the use of alpha features or some other representation of repeated features, but with a recurrent neural network, these elaborations are not necessary.