@inproceedings{ichihara-etal-2025-theoretical,
title = "Theoretical Guarantees for Minimum {B}ayes Risk Decoding",
author = "Ichihara, Yuki and
Jinnai, Yuu and
Ariu, Kaito and
Morimura, Tetsuro and
Uchibe, Eiji",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.acl-long.793/",
doi = "10.18653/v1/2025.acl-long.793",
pages = "16262--16284",
ISBN = "979-8-89176-251-0",
abstract = "Minimum Bayes Risk (MBR) decoding optimizes output selection by maximizing the expected utility value of an underlying human distribution. While prior work has shown the effectiveness of MBR decoding through empirical evaluation, few studies have analytically investigated why the method is effective. As a result of our analysis, we show that, given the size $n$ of the reference hypothesis set used in computation, MBR decoding approaches the optimal solution with high probability at a rate of $\mathcal{O}(n^{-\frac{1}{2}})$, under certain assumptions, even though the language space $\mathcal{Y}$ is significantly larger $|\mathcal{Y}| \gg n$.This result helps to theoretically explain the strong performance observed in several prior empirical studies on MBR decoding. In addition, we provide the performance gap for maximum-a-posteriori (MAP) decoding and compare it to MBR decoding. The result of this paper indicates that MBR decoding tends to converge to the optimal solution faster than MAP decoding in several cases."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ichihara-etal-2025-theoretical">
<titleInfo>
<title>Theoretical Guarantees for Minimum Bayes Risk Decoding</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yuki</namePart>
<namePart type="family">Ichihara</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuu</namePart>
<namePart type="family">Jinnai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kaito</namePart>
<namePart type="family">Ariu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tetsuro</namePart>
<namePart type="family">Morimura</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eiji</namePart>
<namePart type="family">Uchibe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-251-0</identifier>
</relatedItem>
<abstract>Minimum Bayes Risk (MBR) decoding optimizes output selection by maximizing the expected utility value of an underlying human distribution. While prior work has shown the effectiveness of MBR decoding through empirical evaluation, few studies have analytically investigated why the method is effective. As a result of our analysis, we show that, given the size n of the reference hypothesis set used in computation, MBR decoding approaches the optimal solution with high probability at a rate of \mathcalO(n⁻\frac12), under certain assumptions, even though the language space \mathcalY is significantly larger |\mathcalY| \gg n.This result helps to theoretically explain the strong performance observed in several prior empirical studies on MBR decoding. In addition, we provide the performance gap for maximum-a-posteriori (MAP) decoding and compare it to MBR decoding. The result of this paper indicates that MBR decoding tends to converge to the optimal solution faster than MAP decoding in several cases.</abstract>
<identifier type="citekey">ichihara-etal-2025-theoretical</identifier>
<identifier type="doi">10.18653/v1/2025.acl-long.793</identifier>
<location>
<url>https://aclanthology.org/2025.acl-long.793/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>16262</start>
<end>16284</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Theoretical Guarantees for Minimum Bayes Risk Decoding
%A Ichihara, Yuki
%A Jinnai, Yuu
%A Ariu, Kaito
%A Morimura, Tetsuro
%A Uchibe, Eiji
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F ichihara-etal-2025-theoretical
%X Minimum Bayes Risk (MBR) decoding optimizes output selection by maximizing the expected utility value of an underlying human distribution. While prior work has shown the effectiveness of MBR decoding through empirical evaluation, few studies have analytically investigated why the method is effective. As a result of our analysis, we show that, given the size n of the reference hypothesis set used in computation, MBR decoding approaches the optimal solution with high probability at a rate of \mathcalO(n⁻\frac12), under certain assumptions, even though the language space \mathcalY is significantly larger |\mathcalY| \gg n.This result helps to theoretically explain the strong performance observed in several prior empirical studies on MBR decoding. In addition, we provide the performance gap for maximum-a-posteriori (MAP) decoding and compare it to MBR decoding. The result of this paper indicates that MBR decoding tends to converge to the optimal solution faster than MAP decoding in several cases.
%R 10.18653/v1/2025.acl-long.793
%U https://aclanthology.org/2025.acl-long.793/
%U https://doi.org/10.18653/v1/2025.acl-long.793
%P 16262-16284
Markdown (Informal)
[Theoretical Guarantees for Minimum Bayes Risk Decoding](https://aclanthology.org/2025.acl-long.793/) (Ichihara et al., ACL 2025)
ACL
- Yuki Ichihara, Yuu Jinnai, Kaito Ariu, Tetsuro Morimura, and Eiji Uchibe. 2025. Theoretical Guarantees for Minimum Bayes Risk Decoding. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 16262–16284, Vienna, Austria. Association for Computational Linguistics.