[Submitted on 6 Feb 2024 (v1), last revised 10 Nov 2026 (this version, v3)]
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 $O(n^2)$ to $O(n)$, while empirically preserving most of the quality gains of MBR decoding. We release our source code at this https URL
Submission history
From: Jannis Vamvas [view email]
[v1] Tue, 6 Feb 2024 18:59:30 UTC (47 KB)
[v2] Mon, 3 Jun 2024 09:33:14 UTC (49 KB)
[v3] Tue, 10 Nov 2026 08:05:20 UTC (45 KB)
