Computational Approaches to Human Sentence Processing

4 credits · Graduate

The number of possible parses of a 10-word sentence is around the number of stars in the observable universe. How do our minds find the right star? Experimental approaches try to answer narrowly-scoped pieces of this question, but these have long coexisted and interacted with a growing stream of work using computational language processing systems as broad-coverage models of mental language processing systems. Through lectures, primary readings, discussions, and projects, this course will cover the trajectory of progress in computational psycholinguistics over the past few decades, the families of approaches used (symbolic and connectionist), the insights acquired so far, and the current technological and scientific frontiers. This course is an upper-division/graduate course, and prior coursework in natural language processing, natural language syntax, and/or psycholinguistics is required.