The All-Thing

All carrot and no stick, since ought-three.

江南曲 (李益)


| web page

Other views:
RSS 0.92
RSS 1.0 (broken)

Past posts:

Sun Mon Tue Wed Thu Fri Sat
26 27 28 29 30 31  

Recent comments:
/research/surprise.language.results (5 days ago)
/media/music/cantab.virginity (5 days ago)

Recent search referers:
"tentea. com" (x2)
linux dell x300
craigslist uk
bittorrent pr0n (x3)
"named entity extraction" perl code


Creative Commons License
This work is licensed under a Creative Commons License.

Tue, 01 Jul 2003

Surprise Language

For the past month I've been involved in the TIDES Suprise Language project, which is a cool new ideawhere a bunch of computational linguistics researchers get together and have exactly one month to develop various bits of NLP technology (machine translation, named entity extraction, etc) for a given human language. The surprise bit is that they don't know which language it will be until the month begins—the only thing they know is that it will be a language that people haven't put a lot of effort (in terms of NLP) in to before.

Beyond getting people to start working on a new, presumably relevant, language, this is also a great way of assessing the ability of the research community to do rapid start NLP—something the government is very interested in as of late.

The language was Hindi. The month ended yesterday. Overall the project was pretty successful, I think. All the big players were involved. It was interesting to see a cooperation-based project rather than the competition-style events, which is what I've been involved in before, for conferences. There's still competition in terms of who can develop the best system, but resources, tools, etc. were all shared between researchers.

We submitted a named-entity tagger. It'll be interesting to see how it fares.

Posted at 10:18 | /research | (leave a comment) | permalink


Kindness is a language which the deaf can hear and the blind can read. -- Mark Twain