On the Predictive Power of Neural Language Models for Human Real-Time Comprehension Behavior

TL;DR

This study evaluates various neural language models to determine how well their predictions of next words align with human reading times, revealing architecture-dependent differences and the limited role of syntactic knowledge in predicting comprehension behavior.

Contribution

It systematically compares multiple neural language models' ability to predict human reading behavior and examines the relationship between model expectations, architecture, and syntactic knowledge.

Findings

Deep Transformer and n-gram models outperform others in psychometric predictive power.

Better next-word expectations correlate with improved prediction of human reading times.

Syntactic knowledge does not significantly enhance predictive power beyond perplexity levels.

Abstract

Human reading behavior is tuned to the statistics of natural language: the time it takes human subjects to read a word can be predicted from estimates of the word's probability in context. However, it remains an open question what computational architecture best characterizes the expectations deployed in real time by humans that determine the behavioral signatures of reading. Here we test over two dozen models, independently manipulating computational architecture and training dataset size, on how well their next-word expectations predict human reading time behavior on naturalistic text corpora. We find that across model architectures and training dataset sizes the relationship between word log-probability and reading time is (near-)linear. We next evaluate how features of these models determine their psychometric predictive power, or ability to predict human reading behavior. In general, the better a model's next-word expectations, the better its psychometric predictive power. However, we find nontrivial differences across model architectures. For any given perplexity, deep Transformer models and n-gram models generally show superior psychometric predictive power over LSTM or structurally supervised neural models, especially for eye movement data. Finally, we compare models' psychometric predictive power to the depth of their syntactic knowledge, as measured by a battery of syntactic generalization tests developed using methods from controlled psycholinguistic experiments. Once perplexity is controlled for, we find no significant relationship between syntactic knowledge and predictive power. These results suggest that different approaches may be required to best model human real-time language comprehension behavior in naturalistic reading versus behavior for controlled linguistic materials designed for targeted probing of syntactic knowledge.