Unsupervised Transcription of Text and Music

Taylor Berg-Kirkpatrick

UC Berkeley

Tuesday, November 4, 2014
12:30 p.m., Conference Room 5A

A variety of transcription tasks--for example, both historical document transcription and polyphonic music transcription--can be viewed as linguistic decipherment problems. We've investigated an approach to these tasks that involves building a detailed generative model of the relationship between the input (e.g. an image of a historical document) and its transcription (the text the document contains). We've found that these models can be learned in a
completely unsupervised fashion--without ever seeing an example of an input annotated with its transcription--effectively deciphering the hidden correspondence. Using this approach, we've built state-of-the-art systems for both historical document transcription and polyphonic piano music transcription that outperform previous supervised methods for both tasks.

Bio:

Taylor Berg-Kirkpatrick is a graduate student in computer science at the University of California, Berkeley. He works with professor Dan Klein on applying unsupervised learning techniques to natural language problems.  Taylor completed his undergraduate degree in mathematics and computer science at Berkeley as well, where he won the departmental Dorothea Klumpke Roberts Prize in mathematics. As a graduate student, Taylor has received both the Qualcomm Innovation Fellowship and the National Science Foundation Graduate Research Fellowship.