Skip to content
conftrace_
2018 NAACL NAACL 2018

Accelerating NMT Batched Beam Decoding with LMBR Posteriors for Deployment

Abstract

AbstractWe describe a batched beam decoding algorithm for NMT with LMBR n-gram posteriors, showing that LMBR techniques still yield gains on top of the best recently reported results with Transformers. We also discuss acceleration strategies for deployment, and the effect of the beam size and batching on memory and speed.

Authors

Gonzalo Iglesias, William Tambellini, Adrià de Gispert, Eva Hasler, Bill Byrne

Topics

Machine Learning > Optimization & Theory > Optimization Machine Learning > Application Areas > Efficient Computing

Keywords

neural machine translation beam search decoding batched decoding n-gram posterior deployment optimization
Download PDF

More from NAACL 2018

All NAACL 2018 papers →
SystemT: Declarative Text Understanding for Enterprise
Construction of the Literature Graph in Semantic Scholar
Can Neural Machine Translation be Improved with User Feedback?
Pieces of Eight: 8-bit Neural Machine Translation
From dictations to clinical reports using machine translation
Benchmarks and models for entity-oriented polarity detection