One Model for All Domains: Collaborative Domain-Prefix Tuning for Cross-Domain NER

  • 2023-01-28 13:57:47
  • Xiang Chen, Lei Li, Shuofei Qiao, Ningyu Zhang, Chuanqi Tan, Yong Jiang, Fei Huang, Huajun Chen
  • 0


Cross-domain NER is a challenging task to address the low-resource problem inpractical scenarios. Previous typical solutions mainly obtain a NER model bypre-trained language models (PLMs) with data from a rich-resource domain andadapt it to the target domain. Owing to the mismatch issue among entity typesin different domains, previous approaches normally tune all parameters of PLMs,ending up with an entirely new NER model for each domain. Moreover, currentmodels only focus on leveraging knowledge in one general source domain whilefailing to successfully transfer knowledge from multiple sources to the target.To address these issues, we introduce Collaborative Domain-Prefix Tuning forcross-domain NER (CP-NER) based on text-to-text generative PLMs. Specifically,we present text-to-text generation grounding domain-related instructors totransfer knowledge to new domain NER tasks without structural modifications. Weutilize frozen PLMs and conduct collaborative domain-prefix tuning to stimulatethe potential of PLMs to handle NER tasks across various domains. Experimentalresults on the Cross-NER benchmark show that the proposed approach has flexibletransfer ability and performs better on both one-source and multiple-sourcecross-domain NER tasks. Codes will be available in


Quick Read (beta)

loading the full paper ...