Comments on: Where are all legal computational linguistics resources?

By: arafalov

arafalov — Sat, 31 Jan 2009 01:35:02 +0000

I don’t know about machine translation, though it does sound familiar.

Regarding what they do research, there is a couple of high level argumentation and reasoning systems. They use those to teach argumentation and to visualize cases. But that’s too far up the stack for me.

By: I am the worst correspondant in the world

I am the worst correspondant in the world — Fri, 30 Jan 2009 13:46:43 +0000

Don’t they use the parrallel corpora to do machine translation? The idea is that the machine learns one set of inputs (documents in English) emit another set of outputs (documents in, say, French).

Regarding legal corpora – I wonder about the use of machines in the field of litigation. I suspect the reason that it is not used is because lawyers (unlike, say, bankers or even software guys) are cheap and plentiful. Why bother making cheap labour more efficient?

And indeed, why employ specialised, expensive labour (software engineers with a knowledge of natural language processing) to a problem that is solved cheaply every day?

By: arafalov

arafalov — Sat, 17 Jan 2009 17:50:53 +0000

DrNI,

Thanks for the comment. I am aware of Canadian Hansards, EuroParl and JRC-Acquis and will see how they are used. Hopefully, something interesting will turn up that way.

By: DrNI@CLB

DrNI@CLB — Wed, 14 Jan 2009 19:20:49 +0000

You should look into some recent research on parallel corpora. The European Union Government has to translate a lot of legal text into several languages and as far as I know people use these materials as parallel corpora. Of course it should be possible to use them without the parallel option.