so begins a preliminary discussion of translation divergences, such as the lexical-semantic categorial type as in the English I am hungry, in which the predicate hungry is adjectival, compared with the German Ich habe Hunger ('I have hunger'), in which the corresponding Hunger is nominal. Chapters 2 and 3 form the part dealing with the syntactic component. The former discusses the implementation of GB modules coupled with the parameters particular for each of English, Spanish, and German. The latter deals with the two-level morphological processor used in UNITRAN for analysis and generation. Chapter 4, the first in the part dealing with the lexical-semantic component, is to a large extent a variant of Dorr (1993a). It describes the interlingual representation of UNITRAN. The chosen interlingua is an extended version of LCS, used also as a representation of lexical entries. Dorr justifies this choice as follows: [It] is suitable to the task of translating between divergent struct

This article presents methods for biasing statistical translation models to reflect these properties. Evaluation with respect to independent human judgments has confirmed that translation models biased in this fashion are significantly more accurate than a baseline knowledge-free model. This article also shows how a statistical translation model can take advantage of preexisting knowledge that might be available about particular language pairs. Even the simplest kinds of languagespecific knowledge, such as the distinction between content words and function words, are shown to reliably boost translation model performance on some tasks. Statistical models that reflect knowledge about the model domain combine the best of both the rationalist and empiricist paradigms

This thesis proposes a strongly lexicalist approach to machine translation and applies it to the translation of spatial prepositions and prepositional expressions between English and Spanish. Bilingual contrastive knowledge resides solely in the bilingual lexicon and is structured in the form of correspondences between sets of source and target language lexemes related through indices. The resulting architecture maximizes the independence of the monolingual and bilingual components. This independence is demonstrated by developing a grammar of Spanish which is significantly different in its constructions from its analogous English grammar. In particular, relative clauses are analysed through a single rule that allows gaps in subject position, while clitic climbing and doubling are handled through mechanisms not normally found in grammatical descriptions of English. Bilingual lexical rules, in conjunction with the bilingual lexicon, constitute a single, motivated and well defined mechani...

. This paper describes and discusses some theoretical and practical problems arising from developing a system to combine the structured but incomplete information from machine readable dictionaries (MRDs) with the unstructured but more complete information available in corpora for the creation of a bilingual lexical data base, presenting a methodology to integrate information from both sources into a single lexical data structure. The bicord system (BIlingual CORpus-enhanced Dictionaries) involves linking entries in Collins English-French and FrenchEnglish bilingual dictionary with a large English-French and French-English bilingual corpus. We have concentrated on the class of action verbs of movement, building on earlier work on lexical correspondences specific to this verb class between languages (Klavans and Tzoukermann, 1989), (Klavans and Tzoukermann, 1990a), (Klavans and Tzoukermann, 1990b). 1 We first examine the way prototypical verbs of movement are translated in the Collin...

Current question answering systems succeed in many respects regarding questions about textual documents. However, information exists in other media, which provides both opportunities and challenges for question answering. We present results in extending question answering capabilities to video footage captured in a surveillance setting. Our prototype system, called Spot, can answer questions about moving objects that appear within the video. We situate this novel application of vision and language technology within a larger framework designed to integrate language and vision systems under a common representation. We believe that our framework will support the next generation of multimodal natural language information access systems.

Research in machine translation (MT) has resulted in a number of Machine Translation systems that are based on the interlingua approach. This paper reports the results of designing an interlingua for a large-scale, practical MT system designed to translate technical information from English into a number of target languages. After an analysis of the main features of the translation problem faced by this system we describe the principles underlying our design decisions. We address issues such as the design and development methodology, the grain size of the representation, and our efforts to endow the interlingua representation with the ability to degrade gracefully. We conclude that large-scale MT of technical documentation can be achieved with an interlingua-based architecture if the pivotal point of the system, the interlingua, has been designed in a systemsoriented manner. Keywords: Machine Translation, Applications, Knowledge Representation 1 Introduction An interlingual architec...

The paper presents a prototype lexicalist Machine Translation system (based on the so-called `Shake-and-Bake' approach of Whitelock (1992)) consisting of an analysis component, a dynamic bilingual lexicon, and a generation component, and shows how it is applied to a range of MT problems. Multi-Lexeme translations are handled through bi-lexical rules which map bilingual lexical signs into new bilingual lexical signs. It is argued that much translation can be handled by equating translationally equivalent lists of lexical signs, either directly in the bilingual lexicon, or by deriving them through bi-lexical rules. Lexical semantic information organized as Qualia structures (Pustejovsky 1991) is used as a mechanism for restricting the domain of the rules. 1 Introduction Transfer based approaches to machine translation (MT) involve three main phases: analysis, transfer and generation. During analysis, the syntactic and semantic structure of a sentence is made explicit through a source la...

In this paper we discuss the design and development of an interlingua for a large-scale MT project. We also discuss how the resulting KANT interlingua constrains complexity, supports staged development, evolves in a balanced fashion, and seeks maximal coverage. We address issues such as granularity of the data representation, its specification, the types of information it encodes, and how it supports a modular system architecture. Our experience shows that through a careful, reasoned design and implementation effort, it is possible to achieve multilingual target-language generation for extensive technical domains. Keywords: Natural Language Processing, Knowledge Representation, Practical Applications 1 Introduction Varying degrees of success have been achieved by researchers and developers of machine translation (MT) systems. Automatic machine translation of technical text is possible, if the problem is approached the right way. Systems using a knowledge-based approach are usually ch...

An emerging requirement to sift through the increasing flood of text information has led to the rapid development of information filtering technology in the past five years. This study introduces novel approaches for filtering texts regardless of their source language. We begin with a brief description of related developments in text filtering and multilingual information retrieval. We then present three alternative approaches to selecting texts from a multilingual information stream which represent a logical evolution from existing techniques in related disciplines. Finally, a practical automated performance evaluation technique is proposed. 1 Introduction Automatic filtering of information from text sources has become increasingly important as the volume of electronically accessible texts has exploded in recent years. Among these sources of electronically accessible texts are news stories, journal articles and electronic discussion forums. Since it would be impractical for any indiv...

In this paper we present the results of a quantitative evaluation of the discrepancies between the Italian and English lexica in terms of lexical gaps. This evaluation has been carried out in the context of MultiWordNet, an ongoing project that aims at building a multilingual lexical database. The quantitative evaluation of the English-to-Italian lexical gaps shows that the English and Italian lexica are highly comparable and gives empirical support to the MultiWordNet model. 1.