Discriminative ridge regression algorithm for adaptation in statistical machine translation
| Autor: | Germán Sanchis-Trilles, Francisco Casacuberta, Mara Chinea-Rios |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Log-linear model
Statistical machine translation Machine translation Computer science Log-linear weights 02 engineering and technology Translation (geometry) computer.software_genre Machine learning Reduction (complexity) Discriminative model Artificial Intelligence 0202 electrical engineering electronic engineering information engineering Adaptation business.industry 020207 software engineering Regression Pattern recognition (psychology) 020201 artificial intelligence & image processing Computer Vision and Pattern Recognition Artificial intelligence Computational linguistics business LENGUAJES Y SISTEMAS INFORMATICOS computer Discriminative ridge regression algorithm |
| Zdroj: | RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia instname |
| ISSN: | 1433-755X 1433-7541 |
| DOI: | 10.1007/s10044-018-0720-5 |
| Popis: | [EN] We present a simple and reliable method for estimating the log-linear weights of a state-of-the-art machine translation system, which takes advantage of the method known as discriminative ridge regression (DRR). Since inappropriate weight estimations lead to a wide variability of translation quality results, reaching a reliable estimate for such weights is critical for machine translation research. For this reason, a variety of methods have been proposed to reach reasonable estimates. In this paper, we present an algorithmic description and empirical results proving that DRR is able to provide comparable translation quality when compared to state-of-the-art estimation methods [i.e. MERT and MIRA], with a reduction in computational cost. Moreover, the empirical results reported are coherent across different corpora and language pairs. The research leading to these results were partially supported by projects CoMUN-HaT-TIN2015-70924-C2-1-R (MINECO/FEDER) and PROMETEO/2018/004. We also acknowledge NVIDIA for the donation of a GPU used in this work. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink