Computing That Serves

The MSR ESL Assistant: Detecting and Correcting Non-native Errors in English


Thursday, February 19, 2009 - 11:00am


Michael Gamon
Researcher, Microsoft Research

The estimated 750 million non-native speakers of English in the world outnumber native speakers by a large margin. The demand for automatic tools to assist non-native writers is large, and we see this as a great opportunity for the application of NL technology. We present a modular web-based system for detection and correction of errors made by non-native (English as a Second Language = ESL) writers. This system went live in summer of 2008 at, and has generated more than 100k hits. Since then, we have been collecting and analyzing user data. In addition to a technical overview of the system and the methods employed, we will discuss patterns and statistics of system usage.


Michael Gamon has been a researcher in the Natural Language Processing group at Microsoft Research in Redmond, Washington since 1996.  Michael earned his MA (1991) and PhD (1996) in linguistics from the University of Washington.  His current research interests are automatic error correction and writing assistance, social media analysis, and sentiment detection.  Michael has also worked on grammar checking, machine translation, and customer feedback analysis.