Text is data. Language is a window into cognition. By processing language we can extract understanding of the world around us and how our brain processes that world. We can also build useful tools to help people.
My research investigates three fundamental questions:
How do we model the acquisition of textual information and its representation in memory through computational approaches?
How does language provide a window into both basic and higher-order cognitive processes?
How do we assess cognitive processes including language ability, student proficiencies of higher order thinking skills, and clinical deficits?
To address theoretical and methodological issues related to these questions, I conduct basic experiments measuring such factors as memory, comprehension, student writing performance and response times. I also use a variety of cognitive and computational modeling techniques that analyze large-scale complex language and log file data to account for the results and build useful applications. These techniques include machine learning, Deep neural networks, and Natural Language Processing. I incorporate theories and techniques from cognitive psychology, psycholinguistics, computational linguistics, computer science, and education.
From cognitive psychology, neuroscience, and psycholinguistics, I derive a theoretical background for understanding human processing of information and methodologies for investigating and evaluating human responses.
From computational linguistics and computer science, I derive language analysis techniques and machine learning algorithms for modeling language , knowledge, and student behavior.
From education, I integrate pedagogical principals and proficiency frameworks to derive information from which to develop and test theories and apply them in large-scale contexts.
A fundamental approach I use is to model data from what humans do and then train computers to perform similarly. It's not to replace humans, but to provide useful tools that allow the computers to work with humans and make them more effective.
My goal is not just to understand underlying mechanisms through modeling them. It is to develop useful applications that incorporate theories instantiated through technology. These applications permit testing and extending theories of cognitive science, while providing practical applications that have impact in the real-world. What excites me is that methods that I have pioneered are used by millions of student annually to improve student achievement, expand student access, and make learning materials more affordable.