PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Bayesian Adaptation for Statistical Machine Translation
Germán Sanchis-Trilles and Francisco Casacuberta
In: 8th International Workshop on Statistical Pattern Recognition, August 18--20, Cesme, Izmir, Turkey.

Abstract

In many pattern recognition problems, learning from training samples is a process that requires important amounts of training data and a high computational effort. Sometimes, only limited training data and/or limited computational resources are available, but there is also available a previous system trained for a closely related task and with enough training material. This scenario is very frequent in statistical machine translation and adaptation can be a solution to deal with this problem. In this paper, we present an adaptation technique for (state-of-the-art) log-linear modelling based on the well-known Bayesian learning paradigm. This technique has been applied to statistical machine translation and can be easily extended to other pattern recognition areas in which log-linear models are used. We show empirical results in which a small amount of adaptation data is able to improve both the non-adapted system and a system which optimises the above-mentioned weights only on the adaptation set

EPrint Type:Conference or Workshop Item (Talk)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Natural Language Processing
ID Code:7433
Deposited By:Alfons Juan
Deposited On:17 March 2011