谷歌提交新评测模型，提升翻译质量预测与错误定位精度。
MetricX-25 and GemSpanEval: Google Translate Submissions to the WMT25 Evaluation Shared Task
- 用Gemma 3微调，改进输入格式和训练方式，提升评分预测能力。
- 新模型MetricX-25在MQM和ESA评分上显著优于旧版，误差更小。
- 生成式错误定位模型可同时输出错误上下文，避免歧义，适合评估工具开发者。
本文介绍谷歌在统一的WMT25机器翻译评估共享任务中的提交方案。针对质量分数预测子任务，我们推出了新一代MetricX-25，通过改进输入格式和训练协议，并将Gemma 3调整为仅编码器架构，搭配回归头，有效预测MQM与ESA质量得分，显著优于前代模型。针对错误片段检测子任务，我们开发了GemSpanEval，基于Gemma 3的解码器架构，训练其预测错误片段及其严重性与类别。该模型以生成式任务形式输出每个错误片段的上下文，确保识别无歧义。实验表明，其在错误定位性能上可媲美xCOMET这一强基准模型，且具备更强的可解释性。
原文摘要 · Abstract (English)
In this paper, we present our submissions to the unified WMT25 Translation Evaluation Shared Task. For the Quality Score Prediction subtask, we create a new generation of MetricX with improvements in the input format and the training protocol, while for the Error Span Detection subtask we develop a new model, GemSpanEval, trained to predict error spans along with their severities and categories. Both systems are based on the state-of-the-art multilingual open-weights model Gemma 3, fine-tuned on publicly available WMT data. We demonstrate that MetricX-25, adapting Gemma 3 to an encoder-only architecture with a regression head on top, can be trained to effectively predict both MQM and ESA quality scores, and significantly outperforms its predecessor. Our decoder-only GemSpanEval model, on the other hand, we show to be competitive in error span detection with xCOMET, a strong encoder-only sequence-tagging baseline. With error span detection formulated as a generative task, we instruct the model to also output the context for each predicted error span, thus ensuring that error spans are identified unambiguously.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。