Refined Assessment for Translation Evaluation: Rethinking Machine Translation Evaluation in the Era of Human-Level Systems
Summary
该论文批判了传统机器翻译评估方法（MQM、ESA）的局限，提出了RATE（Refined Assessment for Translation Evaluation）协议和由专业语言学家标注的高质量英俄翻译评估数据集，发现SOTA机器翻译系统在准确度上可能已超越人类翻译，但在流畅度上仍存在差距。
View Cached Full Text
Cached at: 10/28/26, 02:04 AM
# Refined Assessment for Translation Evaluation: Rethinking Machine Translation Evaluation in the Era of Human-Level Systems Source: [https://arxiv.org/abs/2610.33749](https://arxiv.org/abs/2610.33749) [View PDF](https://arxiv.org/pdf/2610.33749) > Abstract:As machine translation systems approach human\-level quality, traditional evaluation methodologies struggle to detect subtle translation errors\. We critically examine limitations in current gold\-standard approaches \(MQM and ESA\), including inconsistencies from variable annotator expertise, excessive categorization complexity, coarse severity granularity, accuracy bias over fluency, and time constraints\. To address this issue, we introduce a high\-quality dataset consisting of human evaluations for English\-Russian translations from WMT24, created by professional linguists\. We show that expert assessments without time pressure yield substantially different results from standard evaluations\. To enable consistent and rich annotation by these experts, we developed the RATE \(Refined Assessment for Translation Evaluation\) protocol\. RATE provides a streamlined error taxonomy, expanded severity ratings, and multidimensional scoring balancing accuracy and fluency, facilitating deeper analysis of MT outputs\. Our analysis, powered by this expert dataset, reveals that state\-of\-the\-art MT systems may have surpassed human translations in accuracy while still lagging in fluency \- a critical distinction obscured by existing accuracy\-biased metrics\. Our findings highlight that advancing MT evaluation requires not only better protocols but crucially, high\-quality annotations from skilled linguists\. ## Submission history From: Dmitry Popov \[[view email](https://arxiv.org/show-email/fc2f5807/2610.33749)\] **\[v1\]**Sat, 24 Oct 2026 15:48:47 UTC \(1,334 KB\)
Similar Articles
Beyond Prompts: A Systematic Study of LLM-Based Machine Translation Evaluators
A systematic study of LLM-based machine translation evaluator design — instruction structure, few-shot demonstrations, backbone choice, and multi-step workflows — introducing Orbit, a family of LLM MT evaluators that achieve expert-level error span detection and ranking at roughly 20× lower cost than human annotation.
Large Language Models for Machine Translation Quality Annotation: Humans and Models Are Both Challenged
This paper evaluates how well LLMs can replace human annotators for machine translation quality assessment across MQM and ESA schemes, 70 language pairs, and WMT23/WMT25 data, finding that both humans and LLMs remain unreliable in many settings and proposing human-LLM collaborative annotation pipelines.
Follow-up to my TranslateGemma-12b benchmark post: human reviewers flagged 71% of the segments automated metrics rated clean
A human review of TranslateGemma-12b's translations revealed that 71% of segments rated clean by automated metrics actually contained errors, highlighting significant gaps in metric-only evaluation for multilingual translation quality.
Last Translation Benchmark
The Last Translation Benchmark introduces a live dataset of peer-reviewed, multimodal examples designed to evaluate and break leading machine translation models, with handcrafted verification rules for reliable assessment. It addresses the saturation of current benchmarks and the unreliability of automatic metrics.
Do Evaluation Metrics Detect Errors in Classical Chinese to English Translations?
This paper investigates whether automatic evaluation metrics for machine translation are reliable for Classical Chinese to English translation, using a diagnostic framework based on minimal pairs. It finds all metrics have blind spots, with MetricX24 performing best overall.