Where did it all go wrong? New Tools for Automatic Error Analysis in NLP

Add to your list(s) Download to your calendar using vCal

• Jonathan Kummerfeld
• Friday 02 August 2013, 12:00-13:00
• FW26, Computer Laboratory.

If you have a question about this talk, please contact Tamara Polajnar.

System performance is typically interpreted through metrics, which quantify errors, but do not analyse them. In this talk I will present new tools that automatically analyse system output, giving a linguistically intuitive breakdown of errors.

For parsing I will present results for English and Chinese, showing how error distributions vary between systems, what types of errors are being resolved by re-rankers, and what types are introduced when parsing out-of-domain text. For co-reference resolution I will present some preliminary results, comparing the errors made by systems in a recent shared task.

Jonathan is a PhD candidate at the University of California, Berkeley, advised by Dan Klein

This talk is part of the NLIP Seminar Series series.

This talk is included in these lists:

• All Talks (aka the CURE list)
• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge Forum of Science and Humanities
• Cambridge Language Sciences
• Cambridge talks
• Chris Davis' list
• Department of Computer Science and Technology talks and seminars
• FW26, Computer Laboratory
• Graduate-Seminars
• Guy Emerson's list
• Interested Talks
• Language Sciences for Graduate Students
• NLIP Seminar Series
• PMRFPS's
• Simon Baker's List
• Trust & Technology Initiative - interesting events
• bld31
• ndk22's list
• ob366-ai4er
• rp587

Note that ex-directory lists are not shown.