This paper is based on the work carried out in the framework of the Verbmobil project, which is a limited-domain speech translation task (German-English). In the nal evaluation, the statistical approach was found to perform best among ve competing approaches. In this

istic methods for syntax, just as for a long time McCarthy and Hayes (1969) discouraged exploration of probabilistic methods in Artificial Intelligence. Among his arguments were that: (i) Probabilistic models wrongly mix in world knowledge (New York occurs more in text than Dayton, Ohio, but for no linguistic reason), (ii) Probabilistic models don't model grammaticality (neither Colorless green ideas sleep furiously nor Furiously sleep ideas green colorless have previously been uttered -- and hence must be estimated to have probability zero, Chomsky wrongly assumes -- but the former is grammatical while the latter is not, and (iii) Use of probabilities does not meet the goal of describing the mind-internal I-language as opposed to the observed-in-the-world E-language. This chapter is not meant to be a detailed critique of Chomsky's arguments -- Abney (1996) provides a survey and a rebuttal, and Pereira (2000) has further useful discussion -- but some of these concerns are still importa

This paper gives an overview of our work on statistical machine translation of spoken dialogues, in particular in the framework of the Verbmobil project. The goal of the Verbmobil project is the translation of spoken dialogues in the domains of appointment scheduling and travel planning. Starting with the Bayes decision rule as in speech recognition, we show how the required probability distributions can be structured into three parts: the language model, the alignment model and the lexicon model. We describe the components of the system and report results on the Verbmobil task. The experience obtained in the Verbmobil project, in particular a large-scale end-to-end evaluation, showed that the statistical approach resulted in signicantly lower error rates than three competing translation approaches: the sentence error rate was 29% in comparison with 52% to 62% for the other translation approaches. 1.

This thesis presents the results of a study on information extraction from unrestricted Turkish text using statistical language processing methods. We have successfully applied statistical methods using both the lexical and morphological information to the following tasks: The Turkish Text Deasciifier task aims to convert the ASCII characters in a Turkish text, into the corresponding non-ASCII Turkish characters (i.e., "fi", ";5", "g", "", "", '5", and their upper cases).

This paper gives an overview of our work on statistical machine translation in the framework of the Verbmobil project. The goal of the Verbmobil project is the translation of spoken dialogues in the domains of appointment scheduling and travel planning. Starting with the Bayes decision rule as in speech recognition, we show how the required probability distributions can be structured into three parts: the language model, the alignment model and the lexicon model. We describe the components of the system and report results on the Verbmobil task. The experience obtained in the Verbmobil project, in particular a large-scale end-to-end evaluation, showed that the statistical approach resulted in signicantly lower error rates than three competing translation approaches: the sentence error rate was 29% in comparison with 52% to 62% for the other translation approaches. 1. INTRODUCTION In comparison with written language, speech and especially spontaneous speech poses additional diculties ...

The performance of machine translation technology after 50 years of development leaves much to be desired. There is a high demand for well performing and cheap MT systems for many language pairs and domains, which automatically adapt to rapidly changing terminology. We argue that for successful MT systems it will be crucial to apply data-driven methods, especially statistical machine translation. In addition, it will be very important to establish common test environments. This includes the availability of large parallel training corpora, well defined test corpora and standardized evaluation criteria. Thereby research results can be compared and this will open the possibility for more competition in MT research.

this paper evolved out of discussions with Michael Root over fifteen years ago. Earlier drafts of this paper were read at the Universities of Salzburg, Florence, Bologna, Genova, Palermo, and Venice; and at a conference honoring Professor Quine at Wittenberg College, April, 1992. I want to thank Professor Quine for his comments there. Also, I had some extremely helpful conversations with Paul Boghossian and Barry Loewer during various stages in writing this paper. I would like to thank Bruce Aune, Burt Dreben, Jerry Fodor and especially Roger Gibson for detailed comments on earlier drafts.

killing interest in probabilistic methods for syntax, just as for a long time McCarthy and Hayes (1969) discouraged exploration of probabilistic methods in Articial Intelligence. Among his arguments were that: (i) Probabilistic models wrongly mix in world knowledge (New York occurs more in text than Dayton, Ohio, but for no linguistic reason), (ii) Probabilistic models don't model grammaticality (neither Colorless green ideas sleep furiously or Furiously sleep ideas green colorless have previously been uttered and hence have probability zero, Chomsky assumes but the former is grammatical while the latter is not, and (iii) Use of probabilities does not meet the goal of describing the mind-internal I-language as opposed to the observed-in-the-world E-language. This chapter is not meant to be a detailed critique of Chomsky's arguments Abney (1996) provides a survey and a rebuttal, and Pereira (2000) has useful further discussion but some of these

This paper gives an overview of the stochastic modelling approach to machine translation. Starting with the Bayes decision rule as in pattern classification and speech recognition, we show how the resulting system architecture can be structured into three parts: the language model probability, the string translation model probability and the search procedure that generates the word sequence in the target language. We discuss the properties of the system components and report results on the translation of spoken dialogues in the VERBMOBIL project.

This master thesis aims to investigate a phrase-based approach of Memory-based Machine Translation. This is a form of automatic translation powered by lazy-learning classifiers to translate fragments of the input sentence. A parallel corpus serves as the basis for training such a classifier. In the phrase-based approach the principal component of these fragments is a phrase of arbitrary length. This can be contrasted to prior research in the field in which this component was a single word. A key element in the research is a comparison of three methods of phrase extraction. A new decoder has been developed to deal with the characteristics unique to this approach, and re-assemble the translated fragments into one final translation. This research will show that one of the proposed phrase-extraction methods is capable of outperforming previous word-based approaches, even though this gain is limited and the impact