Multi-domain Multilingual Named Entity Recognition: Revisiting & Grounding the resources issue (original) (raw)

The paper discusses the challenges and methodologies in developing multilingual and multi-domain Language Resources (LRs) for Named Entity Recognition (NERC). The authors describe their ongoing work aimed at harmonizing initially diverse multilingual LRs for effective reuse across various languages and domains such as travel, politics, and health. The results of preliminary evaluations highlight both the successes and limitations encountered in adapting existing NERC systems, emphasizing the necessity for more robust methodologies to improve performance across different contexts.