Integrating Human Cognition with Natural Language ProcessingI will present a line of research which aims to create a new conceptual and computational framework for integrating the scientific study of human cognition with the field of Natural Language Processing (NLP). This framework puts forward the idea that NLP systems and machine learning techniques can be highly instrumental for modeling how humans learn and process language. In a complementary vein, it posits that cognitive theory and behavioral data from human language processing can facilitate the development of NLP systems which can process language effectively. In this talk I will focus on work which develops these ideas using eye movements during reading. I will first demonstrate the informativeness of human gaze by using it to infer the linguistic background of readers and shed light on cross-linguistic influence in language comprehension. I will then present an approach for leveraging eye movements during question answering for improving machine reading comprehension. Finally, I will discuss a real-world application for assessment of linguistic proficiency as an automatic byproduct of ordinary reading, as well as a related analysis of the relation between linguistic proficiency and responsiveness to the linguistic characteristics of the text.

Prof. Yevgeni Berzak
The Technion - Israel Institute of Technology