Thursday, April 21, 2011

Paper Reading #24: Personalized Reading Support for Second-Language Web Documents by Collective Intelligence

Comments: Derek Landini, Aaron Kirkes.
Reference Information:
Title: Personalized Reading Support for Second-Language Web Documents by Collective Intelligence
Authors: Yo Ehara, Nobuyuki Shimizu, Takashi Ninomiya, Hiroshi Nakagawa
Venue: IUI’10, February 7–10, 2010, Hong Kong, China.

Summary: The authors of this paper are working on a system which can analyzes and translate certain phrases in real time. Rather than a traditional translation system, which converts entire pages over to a user's native language, their system analyzes a page to determine what the primary language on the page is, and then it converts any word it cannot detect are written in that language.

The system uses logistic regression to predict which words are written in a second language. The overall results of the system were highly promising. The user feedback they recovered from their test subjects showed that they enjoyed using the system, and that it was overall highly effective.

Dissuasion: This pair's research was incredibly interesting. It seems like this would be a great browser plugin, or perhaps an extension of Google Translate. I hope that this sort of research continues and improves over time.

1 comment:

  1. Have you ever used word glossing in a browser? The only place i've seen it are for those stupid Bing adds or whatever that are really annoying. Never seen it for word definitions.. that would be nice.

    ReplyDelete