Lecture

This tutorial will be split in two parts. In the first part (duration:
approximately 2h30) we will cover the background and state of the art on
quality estimation at different levels of granularity (word, sentence and
document), from features, to quality labels, machine learning techniques
and evaluation. We will also discuss open challenges in the area. In the
second part (duration: approximately 1h30) we will demonstrate QuEst++
(https://github.com/ghpaetzold/questplusplus), our framework for quality
estimation, including how to set up an experiment and run the code, how to
implement new features at different granularity levels, as well as how to
add other machine learning algorithms to the pipeline.

This tutorial will assume attendees are acquainted with machine translation,
general natural language processing, and basic machine learning.

This tutorial is a training event sponsored by project Abu-Matran
"Automatic building of Machine Translation", an FP7-PEOPLE-2012-IAPP
project" (www.abumatran.eu).

We would appreciate if you confirmed your attendance to Mikel L. Forcada
(mlf@ua.es).