Explore the application of the Expectation Maximization (EM) algorithm in resolving pronoun-anaphora in this 1-hour 12-minute lecture by Eugene Charniak from Brown University. Delve into how EM, typically challenging in unsupervised learning scenarios, proves remarkably effective for this specific natural language processing task. Discover how the algorithm successfully determines thousands of parameters, resulting in state-of-the-art performance in identifying antecedents for pronouns. Gain insights from Charniak's extensive experience in language understanding technologies, statistical language learning, and various areas of language processing including parsing, discourse, and anaphora.
EM Works for Pronoun-Anaphora Resolution - 2009
Center for Language & Speech Processing(CLSP), JHU via YouTube
Overview
Syllabus
EM Works for Pronoun-Anaphora Resolution – Eugene Charniak (Brown University) - 2009
Taught by
Center for Language & Speech Processing(CLSP), JHU