Prediction and Memory in Naturalistic Language Processing
Investigating how prediction and working memory influence everyday language processing
What are the moment-by-moment computations that allow us to rapidly comprehend
what others say to us? Evidence has been accumulating for decades that we incrementally
build up meaning representations by (1) storing and retrieving representations in working
memory and (2) predicting future words and meanings. However, we still don’t have a good
handle on how these computations work, how prediction and working memory relate to each
other, or whether the responses we see to decontextualized, constructed laboratory
sentences still happen during everyday language use.
One of my core lines of research is to study the algorithmic structure and degree of
dissociability of prediction and memory effects during naturalistic sentence comprehension.
To do so, I analyze large-scale naturalistic eye-tracking, self-paced reading, and
fMRI data using tools from natural language processing, linguistic theory, and machine
learning.
Related publications
Cur Dir Psych Sci
Similarity of computations across domains does not imply shared implementation: The case of language comprehension
A large-scale study of the effects of word frequency and predictability in naturalistic reading
Shain, Cory
In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers)
2019