Dartmouth Events

Full events calendar

The Role of Computation in Phonological Typology and Learning

Jane Chandlee completed her PhD in Linguistics at the University of Delaware in 2014, specializing in computational phonology. Prior to that, she completed a MFA in Creative Writing at the Pennsylvania State University and a MA in Applied Linguistics at Boston University. She is currently a Visiting Assistant Professor in the Department of Computer Science at Haverford College, outside Philadelphia.

1/30/2017
4:30 pm – 5:30 pm
Reed Hall 108
Intended Audience(s): Public
Categories: Academic Calendar, Lectures & Seminars

Research in computational linguistics has brought about a range of technological developments that many of us interact with in our daily lives. But the study of the computational nature of language itself also provides key insights into what human language really is and how humans successfully learn it at such a young age.

 
Focusing on phonology, in this talk I discuss how categorizing natural language patterns in terms of their computational complexity can 1) account for the range of patterns we see attested in the world’s languages (and for the range of patterns we don’t see), and 2) lead to provably correct algorithms for learning such patterns from positive data alone.
 
More specifically, I will argue that natural language phonology is restricted to subregular classes of formal languages and functions, and these restrictions provide a highly structured hypothesis space that allows a learner to efficiently learn any pattern in the class. I will also explain how such insights into language itself lead to testable hypotheses for improving the efficiency of algorithms that operate on natural language data.
For more information, contact:
Bise Wood Saint Eugene
bise.wood.saint.eugene@dartmouth.edu
Permanent URL to this event: https://linguistics.dartmouth.edu/events/event?event=43149

Events are free and open to the public unless otherwise noted.