Title: Linear-time Minimum Bayes Risk Decoding with Reference Aggregation
Authors: Jannis Vamvas, Rico Sennrich
Published: 6th February 2024 (Tuesday) @ 18:59:30
Link: http://arxiv.org/abs/2402.04251v1
Abstract
Minimum Bayes Risk (MBR) decoding is a text generation technique that has been shown to improve the quality of machine translations, but is expensive, even if a sampling-based approximation is used. Besides requiring a large number of sampled sequences, it requires the pairwise calculation of a utility metric, which has quadratic complexity. In this paper, we propose to approximate pairwise metric scores with scores calculated against aggregated reference representations. This changes the complexity of utility estimation from to , while empirically preserving most of the quality gains of MBR decoding. We release our source code at https://github.com/ZurichNLP/mbr