Parallel corpora for 6 Indian languages created on Mechanical Turk 
==================================================================

This directory contains data sets for Bengali, Hindi, Malayalam,
Tamil, Telugu and Urdu.  Each data set was created by taking around
100 Indian-language Wikipedia pages and obtaining four independent
translations of each of the sentences in those documents.  The
procedure used to create them, along with descriptions of initial
experiments, is described in:

  "Constructing Parallel Corpora for Six Indian Languages via
  Crowdsourcing".  2012.  Matt Post, Chris Callison-Burch, and Miles
  Osborne.  Proceedings of the NAACL Workshop for Statistical Machine
  Translation (WMT).

The corpora are organized into directories by language pairs:

	bn-en/		Bengali-English
	hi-en/		Hindi-English
	ml-en/		Malayalam-English
	ta-en/		Tamil-English
	te-en/		Telugu-English
	ur-en/		Urdu-English

Within each directory, you'll find the following files:

  PAIR/
       PAIR.metadata
       dict.PAIR.{LANG,en}
       training.PAIR.{LANG,en,seg_ids}
       dev.PAIR.{LANG,en.{0,1,2,3},seg_ids}
       devtest.PAIR.{LANG,en.{0,1,2,3},seg_ids}
       test.PAIR.{LANG,en.{0,1,2,3},seg_ids}
       votes.LANG

The metadata file is organized into rows with four columns each.  The
rows correspond to the original documents that were translated, and
the colums denote (1) the (internal) segment ID assigned to the
document (2) the document's original title (3) a translation of the
title (4) the manual category assignment we assigned to the document.

The data splits were constructed by manually assigning the documents
to one of eight categories (Technology, Sex, Language and Culture,
Religion, Places, People, Events, and Things), and then selecting
about 10% of the documents in each category for dev, devtest, and test
data (that is, roughly 30% of the data), and the remaining for
training data.  Corresponding to each split is a file containing the
segment ID of each sentence. The segment ID identifies the original
document ID and the sentence number within that document.  A metadata
file in each directory matches between document IDs, Wikipedia page
name, a corresponding English translation, and the manual
categorization.

The dictionaries were created in a separate MTurk job.  We suggest
that you append them to the end of your training data when you train
the translation model (as was done in the paper).

The votes files contain the results from a separate MTurk task wherein
new Turkers were asked to vote on which of the four translations of a
given sentence was the best.  We have such information for all languages
except Malayalam.  The format of the votes file is:

  seg_id num_votes sentence votes [sentence votes ...]

Since the data was created by non-expert translators hired over
Mechanical Turk, it's of mixed quality.  We are currently researching
ways of improving the quality of the translations that we solicit in
this way.  However, this data should be enough to get you started
training models.  You can download it here:

  http://joshua-decoder.org/indian-parallel-corpora/

In addition, there are some scripts in the scripts/ that manipulate
the data in various ways.