Symbolic-to-statistical hybridization: extending generation-heavy machine translation

Abstract

The last few years have witnessed an increasing interest in hybridizing surface-based statistical approaches and rule-based symbolic approaches to machine translation (MT). Much of that work is focused on extending statistical MT systems with symbolic knowledge and components. In the brand of hybridization discussed here, we go in the opposite direction: adding statistical bilingual components to a symbolic system. Our base system is Generation-heavy machine translation (GHMT), a primarily symbolic asymmetrical approach that addresses the issue of Interlingual MT resource poverty in source-poor/target-rich language pairs by exploiting symbolic and statistical target-language resources. GHMT's statistical components are limited to target-language models, which arguably makes it a simple form of a hybrid system. We extend the hybrid nature of GHMT by adding statistical bilingual components. We also describe the details of retargeting it to Arabic-English MT. The morphological richness of Arabic brings several challenges to the hybridization task. We conduct an extensive evaluation of multiple system variants. Our evaluation shows that this new variant of GHMT-a primarily symbolic system extended with monolingual and bilingual statistical components-has a higher degree of grammaticality than a phrase-based statistical MT system, where grammaticality is measured in terms of correct verb-argument realization and long-distance dependency translation.

Sustainable language technology for African languages

“…Although SMT is the dominant method in MT, it is not self-evident that it is the best method in most cases. When the source language (SL) and target language (TL) are structurally very different, SML seems to encounter serious problems (Habash et al 2009). We see the same in examples below.…”

“…Most often the approach has been to have a statistical approach as a base and the translation result is enhanced with rule-based components (Zbib et al 2012;Nielssen and Ney 2004). Also rule-based systems can be enhanced with statistical components (Habash and Monz 2009). It is claimed that better translation results can be achieved using a hybrid combination of these approaches (Labaka et al 2014).…”

Integrated Approach to Text Analysis in Teaching Translation at Linguistic Department

“…Thus, due to development of computer technologies and various applications modern linguistics is also equipped with computer-assisted methods of research. Text analysis development is not an exception [13][14][15]. The Internet Age promoted greater speed of translation, which can be achieved by using different machine-assisted tools of translation and analysis [1].…”

“…The Internet Age promoted greater speed of translation, which can be achieved by using different machine-assisted tools of translation and analysis [1]. Thus, for example, there are computer tools for analyzing the morphological level of the text, and disclosing categorical divergence in parts of speech distribution typical for a particular Source Language and a Target Language [14]. The use of translation programs makes it possible to make the analysis of structure, meaning, and thematic divergence.…”

Semantic MMT Model Based on Hierarchical Network of Concepts in Chinese-English MT

“…The techniques of hybridization are grouped into three parts: the morphological, lexical, and system level. As an opposite direction, another method is to add statistical bilingual components to a rule-based system [23], which has a higher degree of grammaticality than a phrasebased statistical MT system, where grammaticality is calculated according to correct verb-argument implemented and translation of long-distance dependency.…”