Research

Paper

AI LLM March 13, 2026

Is Human Annotation Necessary? Iterative MBR Distillation for Error Span Detection in Machine Translation

Authors

Boxuan Lyu, Haiyue Song, Zhi Qu

Abstract

Error Span Detection (ESD) is a crucial subtask in Machine Translation (MT) evaluation, aiming to identify the location and severity of translation errors. While fine-tuning models on human-annotated data improves ESD performance, acquiring such data is expensive and prone to inconsistencies among annotators. To address this, we propose a novel self-evolution framework based on Minimum Bayes Risk (MBR) decoding, named Iterative MBR Distillation for ESD, which eliminates the reliance on human annotations by leveraging an off-the-shelf LLM to generate pseudo-labels.Extensive experiments on the WMT Metrics Shared Task datasets demonstrate that models trained solely on these self-generated pseudo-labels outperform both unadapted base model and supervised baselines trained on human annotations at the system and span levels, while maintaining competitive sentence-level performance.

Metadata

arXiv ID: 2603.12983
Provider: ARXIV
Primary Category: cs.CL
Published: 2026-03-13
Fetched: 2026-03-16 06:01

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.12983v1</id>\n    <title>Is Human Annotation Necessary? Iterative MBR Distillation for Error Span Detection in Machine Translation</title>\n    <updated>2026-03-13T13:34:45Z</updated>\n    <link href='https://arxiv.org/abs/2603.12983v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.12983v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Error Span Detection (ESD) is a crucial subtask in Machine Translation (MT) evaluation, aiming to identify the location and severity of translation errors. While fine-tuning models on human-annotated data improves ESD performance, acquiring such data is expensive and prone to inconsistencies among annotators. To address this, we propose a novel self-evolution framework based on Minimum Bayes Risk (MBR) decoding, named Iterative MBR Distillation for ESD, which eliminates the reliance on human annotations by leveraging an off-the-shelf LLM to generate pseudo-labels.Extensive experiments on the WMT Metrics Shared Task datasets demonstrate that models trained solely on these self-generated pseudo-labels outperform both unadapted base model and supervised baselines trained on human annotations at the system and span levels, while maintaining competitive sentence-level performance.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-03-13T13:34:45Z</published>\n    <arxiv:primary_category term='cs.CL'/>\n    <author>\n      <name>Boxuan Lyu</name>\n    </author>\n    <author>\n      <name>Haiyue Song</name>\n    </author>\n    <author>\n      <name>Zhi Qu</name>\n    </author>\n  </entry>"
}