Introduction to algorithms
Contextual Reasoning in Speech-to-Speech Translation
NLP '00 Proceedings of the Second International Conference on Natural Language Processing
Three heads are better than one
ANLC '94 Proceedings of the fourth conference on Applied natural language processing
Ambiguity preserving machine translation using packed representations
COLING '98 Proceedings of the 17th international conference on Computational linguistics - Volume 1
| Hi-index | 0.00 |
Within the machine translation project Verbmobil, translation is performed simultaneously along four independent translation paths, each implementing a different MT strategy. The four competing translations are combined by a selection procedure so as to form a single optimized output for each input utterance. This selection procedure relies on confidence values that are delivered together with each of the alternative translations. Since the confidence values are computed by four independent modules that are fundamentally different from one another, they are not directly comparable and need to be rescaled in order to gain comparative significance. In this paper we describe a machine learning method tailored to overcome this difficulty by using off line human feedback to determine an optimized confidence rescaling scheme.