... enabling this work we trained a multitude of instances of the same phrase-based SMT sys-tem on 30 distinct combinations of language-pairand domain, each with fourteen distinct trainingsets of increasing ... point but lies withinthe range of the set of initial points. However, it does give ameasure of the closeness of the curve fit using only the initialpoints with the gold fit using all the points; ... (xi;ˆθ)]21/2ctwhere S is the set of training datasets, Tcis theset of test datasets for training configuration c,ˆθ is as defined in Eq. 2, N is the total number of combinations of training configurations...
... correlated with human judgments in its rankings of five Chinese-to-English machine translation systems, and in its ability to distinguishbetween human and machine translations. Bleu’scorrelation with ... without any genuine im-provement intranslation quality. In Sections 3.1and 3.2 we examine ways of synthetically produc-ing such variant translations.3.1 Permuting phrasesOne way in which variation ... The offending en-try was unusual in that it was not fully automatic machine translation; instead the entry was aidedby monolingual English speakers selecting amongalternative automatic translations...
... explosion of the number of rules. It is one of the main reasons for an alternative definition of compositionality within the MiMo system. The nature of the definition allows the translation of both ... mono- lingually. The actual working of the component will be shown in this paper by means of a detailed discussion of wh-movement. Introduction Theoretical research as part ofmachinetranslation ... working of the component will be shown by an elab- orate discussion of wh-movement. Finaily, the translation of anaphoric relations will be defined and some idea will be given of the kind of...
... "Language Engineering : The Real Bottleneck of Natural Language Processing", Proceedings of the 12th International Conference on Computational Linguistics. Nirenburg, S. 1987 Machine Translation, ... "Example-based Approach inMachine Translation& quot;, Proceedings of lnfoJapan "90, Part 2: 65-72. Sumita, E. and Iida, H. 1991 "Acceleration of Example-Based Machine Translation& quot;, ... method of integration in 186 EXPERIMENTS AND PROSPECTS OF EXAMPLE-BASED MACHINETRANSLATION Eiichiro SUMITA* and Hitoshi HDA ATR Interpreting Telephony Research Laboratories Sanpeidani, Inuidani,...
... evaluate the consequence of some con-straints, not only in terms of improved perform-ance but also in terms of loss of information. In- deed, some of the constraints specified in the rule exclude ... (S) indet. plur. (Xit'( )) indet. plur. (Xfr'( )) where Xit' = Xfr', translation equivalent Figure 3: Bilingual LFR of indeterminate plurality In this kind of rules ... time-consuming, and one can legitimately ask if machine- learning methods would do the job. The number of LFRs being relatively restrained in producing neologisms, we can say that the effort of manual...
... solving some of the notorious problems inmachinetranslation (MT) such as disambiguation and selection of target language equivalents. In the following, I shall give a rough summary of what ... The inhouse usage criterion covers a number of aspects determined by special requests of the MT user or the firm ordering the translation. This is first of all a question of inhouse terminology. ... robust state offering the necessary tools and methods of language engineering like weighting mechanisms when their system is about to be applied to large volumes of text with distinct sublanguage...
... proceeds "indirectly" by deriving a more-or-less standard analysis of the input, independent of the TL in- volved (but not necessarily of the SL), and then genera- ting TL output ... analysis; (2) translation proceeds after obtaining a complete syntactic analysis of an input, with no signifi- cant semantic analysis; (3) translation proceeds after obtaining a complete ... sensitivity, but since no interpretation can affect the operation of the parser it still enjoys the advantages of context-free opera- tion. Finally, there is a method for scoring, or rating, interpretations;...
... ABSTRACT In this paper we provide an abstract characterisation of different kinds of robust processing inMachineTranslation and Natural Language Processing systems in terms of the kinds of problem ... PROCESSING INMACHINETRANSLATION This paper is an attempt to provide part of the basis for a general theory of robust processing inMachineTranslation (MT) with relevance to other areas of Natural ... well-known technique in systems theory: insuring against the effect of faulty components in crucial parts of a system by computing the result for a given input by a number of different routes....
... as coordinating or subordinating, as a single occurrence or part of an idiomatic frame, etc. Inter- phrasal/inter-clausal behavior of this part of speech is more fully documented in the grammar ... by the insertion of accurate entries in the coding space. When the computer performs an inflection, it also prints a mark in the coding space. 3.2 ENGLISH INFLECTIONS The posteditor inflects ... If, in non-idiomatic combinations, the posteditor wishes to omit translationof an occurrence, he writes the numeral О in the coding space. (Examples of all coding rules are illustrated in...
... of input to a mechanized system; of translation, and of out- put are interpreted in terms of an operational machinetranslation center. The use of machines to do high-volume, high-speed translation ... naive as to think that a machine will be devel- oped which can turn out machinetranslation not re- quiring post-editing. Those of us who have been con- cerned with translationof materials ... were not used in the devel- opment of the MT program. Although I postulate the state of the art ofmachine translation to be of a sufficient level to warrant opera- tional machine translation...
... this way of handling of syntactical ambiguity can be mentioned. First, in the processing of the majority of syntactically ambiguous sentences within an En- glish into Bulgarian translation ... ambiguity problem is shifted from the domain of analysis to the domain of synthesis of sentences. Thus, instead of trying to resolve such ambiguities in the source language (SL), syntac- tically ... substitution of lexemes in the SL with their translational equivalents from the TL results in the resolution of the syn- tactical ambiguity). For instance, in spite of the existence of am- biguous...
... Our interest lies mainly in their role in the domain in terms of the ease in expressing linguistic knowledge required for MT, as well as the ease of implementation in NIT systems. We begin ... purposes. In this paper, we are interested in formalisms which are being used or have applications in the domain of machinetranslation (MT). These can range from specialised languages for linguistic ... specifications for writing SLLPs, as illustrated in Figure 2 (akin to specifications used in software engineering). I linguistic knowledge (in linguistic formalisms) I specifications (in general...
... Josef Och. 2003. Minimum error rate training in statistical machine translation. In Proceedings of the ACL, pages 160–167.Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. BLEU: ... sam-ple of headlines produced by our system, whichincludes both good and not-so-good outputs.5 MachineTranslation usingWIDL-expressionsWe also employ our WIDL-based realization en-gine in a machine ... 9.2MEnglish words of parallel text), trigram lan-guage model trained on 155M words of Englishnewswire, interpolation weights(Equation 2)trained using discriminative training (Och, 2003)(on...
... YangDepartment of Electrical EngineeringUniversity of Washin g tonSeattle, WA, USAyangmei@ee.washington.eduKatrin KirchhoffDepartment of Electrical EngineeringUniversity of Washin g tonSeattle, ... cross-language information re-trieval. More recently, (Xi and Hwa, 2005) haveused backoff models for combining in- domain and42Phrase-Based Backoff Models for MachineTranslationof Highly In ectedLanguagesMei ... when training data is limited or froma different domain. Many current and future ap-plications ofmachinetranslation require the rapidporting of existing systems to new languages anddomains...