This paper describes a new model-based speaker adaptation algorithm called the eigenvoice approach. The approach constrains the adapted model to be a linear combination of a small number of basis vectors obtained offline from a set of reference speakers, and thus greatly reduces the number of free parameters to be estimated from adaptation data. These "eigenvoice" basis vectors are orthogonal to each other and guaranteed to represent the most important components of variation between the reference speakers. Experimental results for a small-vocabulary task (letter recognition) given in the paper show that the approach yields major improvements in performance for tiny amounts of adaptation data. For instance, we obtained 16% relative improvement in error rate with one letter of supervised adaptation data, and 26% relative improvement with four letters of supervised adaptation data. After a comparison of the eigenvoice approach with other speaker adaptation algorithms, the paper concludes with a discussion of future work.

Speech recognition systems must often decide between competing ways of breaking up the acoustic input into strings of words. Since the possible strings may be acoustically similar, a language model is required; given a word string, the model returns its linguistic probability. This thesis discusses several Markov language models. Subsequently, we present a new kind of language model which • Acknowledgements

This article describes a machine translation system based on an automatic post-editing strategy: initially translate the input text into the target-language using a rule-based MT system, then automatically post-edit the output using a statistical phrase-based system. An implementation of this approach based on the SYSTRAN and PORTAGE MT systems was used in the shared task of the Second Workshop on Statistical Machine Translation. Experimental results on the test data of the previous campaign are presented.

We describe a mixture-model approach to adapting a Statistical Machine Translation System for new domains, using weights that depend on text distances to mixture components. We investigate a number of variants on this approach, including cross-domain versus dynamic adaptation; linear versus loglinear mixtures; language and translation model adaptation; different methods of assigning weights; and granularity of the source unit being adapted to. The best methods achieve gains of approximately one BLEU percentage point over a state-of-the art non-adapted baseline system.

We discuss different strategies for smoothing the phrasetable in Statistical MT, and give results over a range of translation settings. We show that any type of smoothing is a better idea than the relativefrequency estimates that are often used. The best smoothing techniques yield consistent gains of approximately 1% (absolute) according to the BLEU metric.