A Comprehensive Comparison of Neural Networks as Cognitive Models of Inflection

TL;DR

This paper compares various neural network architectures to human judgments on morphological inflection tasks, finding Transformers align more closely with human behavior than LSTMs, challenging assumptions about accuracy and human-likeness.

Contribution

It provides a comprehensive comparison of neural network models as cognitive accounts of inflection, highlighting the Transformer’s superior alignment with human judgments over LSTMs.

Findings

Transformers better match human judgments than LSTMs.

LSTM features improving accuracy do not always enhance human-likeness.

Neural network architecture influences cognitive plausibility in inflection tasks.

Abstract

Neural networks have long been at the center of a debate around the cognitive mechanism by which humans process inflectional morphology. This debate has gravitated into NLP by way of the question: Are neural networks a feasible account for human behavior in morphological inflection? We address that question by measuring the correlation between human judgments and neural network probabilities for unknown word inflections. We test a larger range of architectures than previously studied on two important tasks for the cognitive processing debate: English past tense, and German number inflection. We find evidence that the Transformer may be a better account of human behavior than LSTMs on these datasets, and that LSTM features known to increase inflection accuracy do not always result in more human-like behavior.