The phrase-based translation system Moses has been extended to take advantage of multi-core systems by using multi-threaded decoding. This paper describes how these extensions were implemented and how they can be used, as well as offering some experimental measurements of the potential speed-ups available. Details are also provided of how the multi-threaded Moses library is used to create the Moses server, a platform for building online translation systems.
We describe an open-source implementation of the Margin Infused Relaxed Algorithm (MIRA) for statistical machine translation (SMT). The implementation is part of the Moses toolkit and can be used as an alternative to standard minimum error rate training (MERT). A description of the implementation and its usage on core feature sets as well as large, sparse feature sets is given and we report experimental results comparing the performance of MIRA with MERT in terms of translation quality and stability.
Nicola Bertoldi, Barry Haddow and Jean-Baptiste Fouet
Improved Minimum Error Rate Training in Moses
We describe an open-source software for minimum error rate training (MERT) for statistical machine translation (SMT). This was implemented within the Moses toolkit, although it is essentially standsalone, with the aim of replacing the existing implementation with a cleaner, more flexible design, in order to facilitate further research in weight optimisation. A description of the design is given, as well as experiments to compare performance with the previous implementation and to demonstrate extensibility.