In this project we explore how to use NLP and eye movements in reading to determine various properties of texts, such as their readability level and how engaging they are. We further investigate how to personalize readability and content in real time, with an emphasis on language learners and people with cognitive impairments.
We are developing computational frameworks for decoding linguistic knowledge and cognitive state from eye movements during reading. In particular, we are developing a new type of language assessment technologies where language proficiency is determined as an automatic byproduct of ordinary reading.
In this project we investigate how to bring NLP closer to human language processing abilities by providing NLP systems with inductive biases from human eye movements in reading and brain activity during language comprehension. Example tasks include machine reading comprehension and language modeling.
