MetaMetrics-MT: Tuning Meta-Metrics for Machine Translation via Human Preference Calibration

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Anugraha, David, Kuwanto, Garry, Susanto, Lucky, Wijaya, Derry Tanti, Winata, Genta Indra
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910680542609408
author Anugraha, David
Kuwanto, Garry
Susanto, Lucky
Wijaya, Derry Tanti
Winata, Genta Indra
author_facet Anugraha, David
Kuwanto, Garry
Susanto, Lucky
Wijaya, Derry Tanti
Winata, Genta Indra
contents We present MetaMetrics-MT, an innovative metric designed to evaluate machine translation (MT) tasks by aligning closely with human preferences through Bayesian optimization with Gaussian Processes. MetaMetrics-MT enhances existing MT metrics by optimizing their correlation with human judgments. Our experiments on the WMT24 metric shared task dataset demonstrate that MetaMetrics-MT outperforms all existing baselines, setting a new benchmark for state-of-the-art performance in the reference-based setting. Furthermore, it achieves comparable results to leading metrics in the reference-free setting, offering greater efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2411_00390
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MetaMetrics-MT: Tuning Meta-Metrics for Machine Translation via Human Preference Calibration
Anugraha, David
Kuwanto, Garry
Susanto, Lucky
Wijaya, Derry Tanti
Winata, Genta Indra
Computation and Language
Artificial Intelligence
Machine Learning
We present MetaMetrics-MT, an innovative metric designed to evaluate machine translation (MT) tasks by aligning closely with human preferences through Bayesian optimization with Gaussian Processes. MetaMetrics-MT enhances existing MT metrics by optimizing their correlation with human judgments. Our experiments on the WMT24 metric shared task dataset demonstrate that MetaMetrics-MT outperforms all existing baselines, setting a new benchmark for state-of-the-art performance in the reference-based setting. Furthermore, it achieves comparable results to leading metrics in the reference-free setting, offering greater efficiency.
title MetaMetrics-MT: Tuning Meta-Metrics for Machine Translation via Human Preference Calibration
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2411.00390