System Combination Using Discriminative Cross-Adaptation

Cross-adaptation (CA) based methods of machine translation (MT) system combination work by adapting the decoding step of a baseline system using information from alternate systems. Generally, the required information is very deep, such as a full decoding forest. In this paper, we describe a method of cross-adaptation based system combination which only requires the final output from each alternate system. This is achieved by adding a discriminatively weighted n-gram confidence feature to our decoder. In order to optimize the confidence weight of each system, we present a novel procedure called non-linear Expected-BLEU optimization that can be used to optimize arbitrary nonlinear parameters for any decoding feature. We also describe a method for explicitly creating an adapted system that is dissimilar from each particular input system, which we have found to be useful in combination. Although our new method does not outperform a state-of-the-art confusion network (CN) based combination system on its own, we obtain statistically significant gains of 0.21-0.45 BLEU when the CA output is used as an additional system in CN combination. ∗This work was supported by DARPA/I2O Contract No. HR0011-06-C-0022 under the GALE program (Approved for Public Release, Distribution Unlimited). The views, opinions, and/or findings contained in this article/presentation are those of the author/presenter and should not be interpreted as representing the official views or policies, either expressed or implied, of the Defense Advanced Research Projects Agency or the Department of Defense.