Monolingual Corpus-based MT using Chunks

In this article, we present an automated approach of extracting English-Bengali parallel fragments of text from comparable corpora created using Wikipedia documents. Our approach exploits the multilingualism of Wikipedia. The most important fact is that this approach does not need any domain specific corpus. We have been able to improve the BLEU score of an existing domain specific English-Bengali machine translation system by 11.14%.

2009

Statistical MT is limited by reliance on large parallel corpora. We propose Lemmatic MT, a new paradigm that extends MT to a far broader set of languages, but requires substantial manual encoding effort.

Explorations across Languages and Corpora, 2011

Machine Translation applications have achieved significant results for specialised discourse (cf. Wilks, 2009). The conjunction of web-based resources and hybrid systems seems to render some note-worthy advances. In addition, a novel technique outlines a context approach MT that may be coming into its own (cf. Carbonell et al., 2006). This context-based method focuses on the management of a massive target corpus alone (i.e., processes the translated data from a bilingual dictionary in context). Our own project applies these techniques to the translation of English to Spanish, and attempts to manage and decode n-grams from the source text to the Spanish corpus, via their integration in databases. The process requires as much context (text) as possible and may be partially tested by working with a small specific corpus. The n-gram classification should enable the step called 'flooding' for the scoring of the best translation options.

Journal for Language Technology and Computational Linguistics

METIS-II was a EU-FET MT project running from October 2004 to September 2007, which aimed at translating free text input without resorting to parallel corpora. The idea was to use 'basic' linguistic tools and representations and to link them with patterns and statistics from the monolingual target-language corpus. The METIS-II project has four partners, translating from their 'home' languages Greek, Dutch, German, and Spanish into English. The paper outlines the basic ideas of the project, their implementation, the resources used, and the results obtained. It emphazises on the German implementation. 2 Introduction Starting in October 2004, METIS-II was the continuation of METIS-I (IST-2001-32775) Dologlou et al. (2003). Like METIS-I, METIS-II aims at translating free text input by taking advantage of a combination of statistical, pattern-matching and rule-based methods. The METIS-II project has four partners, each translating from their 'home' languages Greek, Dutch, German, and Spanish into English. The following goals and premises were defined for the project: 1. use 'basic' NLP tools and resources, 2. use bilingual handmade dictionaries, 3. use a monolingual target-language corpus, 4. use translation units within the sentence boundary, 5. allow different tag sets for SL and TL possible, Crucially, parallel corpora are not required, and their usage was excluded within METIS-II. The rationale behind this was to develop prototypes of MT systems which would be suitable to translate 'small languages', i.e. language pairs for which parallel texts are difficult to come by. A basic set of NLP tools is nonetheless required for these languages, albeit very basic. The availability of the monolingual target language corpus,

Advances in Artificial …, 2006

The innovative feature of the system presented in this paper is the use of pattern-matching techniques to retrieve translations resulting in a flexible, language-independent approach, which employs a limited amount of explicit a priori linguistic knowledge. Furthermore, while all state-of-the-art corpus-based approaches to Machine Translation (MT) rely on bitexts, this system relies on extensive target language monolingual corpora. The translation process distinguishes three phases: 1) pre-processing with 'light' rule and statisticsbased NLP techniques 2) search & retrieval, 3) synthesising. At Phase 1, the source language sentence is mapped onto a lemma-to-lemma translated string. This string then forms the input to the search algorithm, which retrieves similar sentences from the corpus (Phase 2). This retrieval process is performed iteratively at increasing levels of detail, until the best match is detected. The best retrieved sentence is sent to the synthesising algorithm (Phase 3), which handles phenomena such as agreement.

2004

Parallel texts are an important resource for applications in multilingual natural language processing and human language technology. This paper presents a method for exploiting available parallel texts, both human translated and revised machine translated texts in order to populate machine translation and translation memory databases.

Proceedings of the sixth conference on Applied natural language processing -, 2000

In this paper we describe an implemented framework for developing monolingual or multilingual natural language generation (NLG) applications and machine translation (MT) applications. The framework demonstrates a uniform approach to generation and transfer based on declarative lexico-structural transformations of dependency structures of syntactic or conceptual levels ("uniform lexico-structural processing"). We describe how this framework has been used in practical NLG and MT applications, and report the lessons learned.

The present article provides a comprehensive review of the work carried out on developing PRESEMT, a hybrid language-independent machine translation (MT) methodology. This methodology has been designed to facilitate rapid creation of MT systems for unconstrained language pairs, setting the lowest possible requirements on specialised resources and tools. Given the limited availability of resources for many languages, only a very small bilingual corpus is required, while language modelling is performed by sampling a large target language (TL) monolingual corpus. The article summarises implementation decisions, using the Greek-English language pair as a test case. Evaluation results are reported, for both objective and subjective metrics. Finally, main error sources are identified and directions are described to improve this hybrid MT methodology.

The main goal of the OpenMT project is the development of open source machine translation architectures based on hybrid models and advanced syntactic-semantic processors. These architectures combine the three main Machine Translation (MT) frameworks, Rule-based (RBMT), Statistical (SMT) and Example-based (EBMT), into hybrid systems. Defined architectures and results will be open source, allow for a rapid development and adaptation of new advanced machine translation systems for other languages. The project deals with four different languages: English, Spanish, Catalan and Basque. The translation systems developed for all language pairs will be internally evaluated using a rich set of linguistic metrics, and in different international evaluation campaigns.

The current paper evaluates the performance of the PRESEMT methodology, which facilitates the creation of machine translation (MT) systems for different language pairs. This methodology aims to develop a hybrid MT system that extracts translation information from large, predominantly monolingual corpora, using pattern recognition techniques. PRESEMT has been designed to have the lowest possible requirements on specialised resources and tools, given that for many languages (especially less widely used ones) only limited linguistic resources are available. In PRESEMT, the main translation process is divided into two phases, the first determining the overall structure of a target language (TL) sentence, and the second disambiguating between alternative translations for words or phrases and establishing local word order. This paper describes the latest version of the system and evaluates its translation accuracy, while also benchmarking the PRESEMT performance by comparing it with other established MT systems using objective measures.