CorPipe 在 CRAC 2025：评估多语言编码器用于多语言共指消解
CorPipe at CRAC 2025: Evaluating Multilingual Encoders for Multilingual Coreference Resolution
AI总结：
CorPipe 25 通过从 TensorFlow 完全重实现为 PyTorch，在 CRAC 2025 多语言共指消解共享任务的 LLM 与无约束赛道中均以 8 个百分点优势夺冠。
摘要：
我们介绍 CorPipe 25，它是 CRAC 2025 多语言共指消解共享任务的获胜系统。该共享任务的第四届在原有无约束赛道之外新增了 LLM 赛道，缩减了开发和测试集规模以降低计算要求，并纳入了更多数据集。CorPipe 25 对我们之前的系统进行了完全重实现，从 TensorFlow 迁移到 PyTorch。我们的系统在 LLM 和无约束两条赛道中均以 8 个百分点的显著优势大幅超越所有其他提交。源代码和训练好的模型已公开，地址为 https://github.com/ufal/crac2025-corpipe。
英文摘要：
We present CorPipe 25, the winning entry to the CRAC 2025 Shared Task on Multilingual Coreference Resolution. This fourth iteration of the shared task introduces a new LLM track alongside the original unconstrained track, features reduced development and test sets to lower computational requirements, and includes additional datasets. CorPipe 25 represents a complete reimplementation of our previous systems, migrating from TensorFlow to PyTorch. Our system significantly outperforms all other submissions in both the LLM and unconstrained tracks by a substantial margin of 8 percentage points. The source code and trained models are publicly available at https://github.com/ufal/crac2025-corpipe.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
