Margin Infused Relaxed Algorithm for Moses

Eva Hasler, Barry Haddow, Philipp Koehn

Research output: Contribution to journalArticlepeer-review

Abstract / Description of output

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.
Original languageEnglish
Pages (from-to)69-78
Number of pages10
JournalPrague Bulletin of Mathematical Linguistics
Publication statusPublished - 2011


Dive into the research topics of 'Margin Infused Relaxed Algorithm for Moses'. Together they form a unique fingerprint.

Cite this