[Corpora-List] FIRE Legal track - 1st Call for participation

Sat Sep 21 17:58:22 UTC 2013

Apologies for cross posting

------------------------------**------------------------------**
-------------------
1st Call for Participation
------------------------------**------------------------------**
-------------------

A pilot track on Information Access in the Legal Domain
To be conducted at FIRE-2013
4 - 6 December 2013, New Delhi, India
http://www.isical.ac.in/~fire/legal.html

------------------------------**------------------------------**
-------------------

The track is divided into two major tasks:

1) *Adhoc-retrieval from legal documents: *
*
*
The task will focus on retrieval of legal documents relevant to a given
situation.

Many forums are popping up each day where people ask for legal advice given
a particular scenario.

e.g. I was driving from work to home last night when a car hit my bike.
Though I was not injured my bike was damaged badly. Insurance company says
I cannot claim insurance unless I file a F.I.R. But I do not want to get
into legal mess. Is there a provision where I can claim an insurance
without filing a F.I.R.

Currently these questions are answered by legal experts or other people who
have at some point of time required the same advice.

The aim of this task is to automatically return documents that can answer
such queries.

Thinking of another another scenario. It can be a lawyer wanting to know
precedents while defending his client.
e.g. Cases where an accused of hit-and-run charge was acquitted.

The aim of this task is to build a system that takes as input the required
needs of a common man/expert and returns relevant legal documents.

This year we will be focusing only on the two most popular domains:
a) Consumer Law
b) Hindu Marriage and Divorce Laws

Corpus will consist of verdicts from various high courts and the supreme
court of India, Indian constitution and various bills and acts of the
Indian parliament.

*2) Identification and Classification of Propositions in Court Judgments*

This task is designed so as to facilitate task 1. But this year both of
these are totally independent. In this task the participants take a legal
document as input and the output will be a parsed document with each
proposition segmented and annotated separately and classified into
different categories.

The flow will be like below:

1) Break a given document into sentences (The task would not be as simple
as identifying full stops to separate sentences as there maybe hyperlinks
or abbreviations in the documents)
2) Identify if a sentence is composed of multiple propositions and break it
down into individual propositions
3) Classify each of the propositions into one of the classes mentioned at
http://www.isical.ac.in/~fire/legal.html#Scheme

Here proposition stands for a smallest part of the sentence that is
meaningful (In most cases separating simple sentences from a given complex
sentence will do the task)

For example if the given sentence is as below:

Some time in November 2012, the accused broke off the relationship, and
became engaged to someone else.

the expected output is:

a. Sometime in November 2012 the accused broke off the relationship
b. and he became engaged to someone else

In order to simplify the task we have decided to allow participants to
focus on either segmentation or classification.

For the segmentation part we will provide the participants with ~800
un-parsed documents which comprise of the past judgments of the Supreme
court of India.
Apart from this we will provide a training set of ~20 parsed documents with
each propositions segmented and annotated by legal experts.

For task 2 the participants need to submit their systems which shall be
evaluated by us on a test corpus.

------------------------------**------------------------------**
-------------------
Important Dates
------------------------------**------------------------------**
-------------------
20 September, 2013     Release of training corpus
20 October, 2013         Submission of working systems
15 Nov, 2013                Release of qrels (result notification)
20 Nov, 2013                Working notes due
------------------------------**------------------------------**
-------------------

Contact:

For the latest updates please join the google mailing group
Information Access in Legal
Domain<https://groups.google.com/forum/#!forum/legaltrack>

After joining the Group the participants can simply mail their doubts to
legaltrack at googlegroups.com

Alternately you can email the queries to:

Parth Mehta
parth.mehta126 at gmail.com

Kripabandhu Ghosh
kripa.ghosh at gmail.com

Madhulika Agrawal
madhu1agrawal at gmail.com

-- 
Regards,
Parth Mehta
DA-IICT
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20130921/41574d50/attachment.htm>
-------------- next part --------------
_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora