This paper describes our recent work on integrating speech recognition and machine translation for improving speech translation performance. Two approaches are applied and their performance are evaluated in the workshop of IWSLT 2005. The first is direct N-best hypothesis translation, and the second, a pseudo-lattice decoding algorithm for translating word lattice, can dramatically reduce compu...

This paper presents the results obtained by the University at Buffalo (UB) in CLEF 2003. Our efforts concentrated in the monolingual retrieval and large multilingual retrieval tasks. We used a modified version of the SMART system, a heuristic method based on bigrams to generate phrases that works across multiple languages, and pseudo relevance feedback. Query translation was performed using pub...

In this paper, we identify factors that affect machine translation (MT) of a source query for cross-language information retrieval (CLIR) and empirically evaluate the effect of pseudo relevance feedback on crosslanguage retrieval performance. Our experiments demonstrate that, by using pseudo relevance feedback, we can significantly improve cross-language retrieval performance and achieve the le...

Abstract We present a natural standard translation of inquisitive modal logic $\mathrm{InqML}$ into first-order over the two-sorted relational representations intended models, which captures built-in higher-order features . This is based on graded notion flatness that ties inherent second-order, team-semantic information states to subsets or tuples bounded size. A pseudo-models, relaxes non-ele...

This paper presents a system overview of an English to Hindi Machine-Aided Translation System named AnglaHindi. Its beta-version has been made available on the internet for free translation at http://anglahindi.iitk.ac.in AnglaHindi is an English to Hindi version of the ANGLABHARTI translation methodology developed by the author for translation from English to all Indian languages. Anglabharti ...

This paper investigates parameter adaptation in Statistical Machine Translation(SMT). To overcome the parameter bias-estimation problem with Minimum Error Rate Training(MERT), we extend it under a transductive learning framework, by iteratively re-estimating the parameters using both development and test data, in which the translation hypotheses of the test data are used as pseudo references. F...

Recent works have shown that synthetic parallel data automatically generated by existing translation models can be an effective solution to various neural machine translation (NMT) issues. In this study, we construct NMT systems using only synthetic parallel data. As an effective alternative to real parallel data, we also present a new type of synthetic parallel corpus. The proposed pseudo para...

Word Sense Disambiguation (WSD) is one of the key issues in natural language processing. Currently, supervised WSD methods are effective ways to solve the ambiguity problem. However, due to lacking of large-scale training data, they cannot achieve satisfactory results. In this paper, we present a WSD method based on context translation. The method is based on the assumption that translation und...

Cross-Language Spoken Document Retrieval (CLSDR) combines both the complexities of retrieval from collections characterized by speech transcription errors and language translation issues between search requests and documents. Thus achieving effective retrieval in this domain is potentially very challenging. For the CLEF 2003 SDR task we adopted a standard query translation strategy using commer...