Multilingual Language Models Predict Human Reading Behavior

04/12/2021
by   Nora Hollenstein, et al.
0

We analyze if large language models are able to predict patterns of human reading behavior. We compare the performance of language-specific and multilingual pretrained transformer models to predict reading time measures reflecting natural human sentence processing on Dutch, English, German, and Russian texts. This results in accurate models of human reading behavior, which indicates that transformer models implicitly encode relative importance in language in a way that is comparable to human processing mechanisms. We find that BERT and XLM models successfully predict a range of eye tracking features. In a series of experiments, we analyze the cross-domain and cross-language abilities of these models and show how they reflect human sentence processing.

READ FULL TEXT
research
06/07/2021

Relative Importance in Sentence Processing

Determining the relative importance of the elements in a sentence is a k...
research
05/19/2020

Comparing Transformers and RNNs on predicting human sentence processing data

Recurrent neural networks (RNNs) have long been an architecture of inter...
research
08/30/2022

Do language models make human-like predictions about the coreferents of Italian anaphoric zero pronouns?

Some languages allow arguments to be omitted in certain contexts. Yet hu...
research
12/23/2022

Why Does Surprisal From Larger Transformer-Based Language Models Provide a Poorer Fit to Human Reading Times?

This work presents a detailed linguistic analysis into why larger Transf...
research
10/09/2021

Leveraging recent advances in Pre-Trained Language Models forEye-Tracking Prediction

Cognitively inspired Natural Language Pro-cessing uses human-derived beh...

Please sign up or login with your details

Forgot password? Click here to reset