提出小型通用多语言命名实体识别模型Otter，性能接近大模型。
What Matters When Building Universal Multilingual Named Entity Recognition Models?
- 系统评估多种模型组件对多语言命名实体识别的影响
- Otter模型在零样本场景下比同类小模型高5.7个百分点F1
- 体积仅大模型的1/90，推理速度快一到两个数量级
近年来，通用多语言命名实体识别（NER）的发展得益于多语言Transformer模型、特定任务架构、自定义损失函数和大规模训练数据集。然而，尽管已有大量工作，我们发现许多模型设计决策缺乏系统性依据，各组件通常在组合中被评估而非单独测试。这阻碍了领域进展，难以确定哪些选择能提升多语言泛化能力。本文对Transformer主干、架构、训练目标、数据构成及阈值选择进行了全面实证评估。基于这些发现，我们提出了Otter——一个小型编码器模型，在零样本多语言NER任务中持续优于强基准模型，相比同等规模模型提升5.7个百分点的F1分数。此外，其性能仅比Qwen3-32B低1.4分，比Gemma3-27B低5.4分，但模型大小约为后者的1/90，推理速度提升一到两个数量级。
原文摘要 · Abstract (English)
Recent progress in universal multilingual named entity recognition (NER) has been driven by multilingual transformer models, task-specific architectures, custom loss functions, and large-scale training datasets. However, despite substantial prior work, we find that many design decisions for such models are made without systematic justification, with individual components evaluated only in combination rather than in isolation. We argue that this impedes progress in the field by making it difficult to identify which choices improve multilingual generalization. In this work, we conduct an extensive empirical evaluation on transformer backbones, architectures, training objectives, data composition, and threshold selection for zero-shot, multilingual NER models. Building on these findings, we present Otter, a small encoder model that achieves consistent improvements over strong multilingual NER baselines, outperforming similarly sized models by 5.7 percentage points in F1. Further, it remains within 1.4 points of Qwen3-32B and 5.4 points below Gemma3-27B, while being roughly 90x smaller and one to two orders of magnitude faster at inference.
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。