MMT: a new machine translation technology for CAT tools
Luisa Bentivogli and Marcello Federico, Trento, Italy
Nowadays, computer-assisted translation (CAT) tools represent the dominant technology in the translation market – and those including machine translation (MT) engines are on the increase. In this new scenario, where MT and post-editing are becoming the standard portfolio for professional translators, it is of the utmost importance that MT systems are specifically tailored to translators.
In this talk, we will present MMT, a new open-source MT software, the development of which has been funded by the European Union. MMT targets two use cases: translation companies that need to install their own MT server using their own data; and professional translators working with CAT tools (currently MMT is accessible in SDL Trados and MateCat). In this presentation, we will focus on the second use.
First, we will introduce MMT’s most distinguishing features when used through a CAT tool: (i) MMT does not require any initial training: as soon as translators upload their translation memories in the CAT tool, MMT seamlessly and quickly learns from this data; (ii) MMT adapts to the content to be translated in real time: the system leverages the training data most similar to the document being translated; (iii) MMT learns from user corrections: during the translation workflow, MMT constantly learns from the post-edited sentences to improve its translation suggestions. Second, we will demonstrate MMT within MateCat, a popular online professional CAT tool.
With this presentation, METM participants will learn about industry trends aiming to develop MT focusing on translators’ specific needs. They will see how current state-of-the-art MT technology is being consolidated into a single, easy-to-use product capable of learning from – and evolving through – interaction with users, with the final aim of increasing MT-output utility for the translator in a real professional environment.
Luisa Bentivogli works as a researcher in the HLT-MT research unit at Fondazione Bruno Kessler, Trento, Italy. Her research activities focus on the evaluation of human language technologies, computational lexicography in a multilingual environment and contrastive linguistics. She contributed to various EU-funded projects, including MateCat and MMT. She has been involved in organizing several evaluation campaigns to assess the state of the art of human language technologies and other international events, including the School of Advanced Technologies for Translators (SATT 2016).
Marcello Federico heads the HLT-MT research unit at Fondazione Bruno Kessler, Trento, Italy. He is also a lecturer and board member at the ICT Doctorate School of the University of Trento, Italy, and scientific advisor to MateCat Srl. His research expertise is in statistical machine translation, spoken language translation, statistical language modelling, information retrieval and speech recognition. He is currently senior editor for the journal IEEE/ACM Transactions on Audio, Speech and Language Processing, and he is a senior member of the IEEE and the ACM.