One Model for All Domains: Collaborative Domain-Prefix Tuning for Cross-Domain NER
This work addresses the low-resource challenge in practical NER applications by enabling flexible knowledge transfer across domains without retraining entire models, though it is incremental as it builds on existing prefix tuning and text-to-text generation techniques.
The paper tackles the problem of cross-domain named entity recognition (NER) by proposing Collaborative Domain-Prefix Tuning (CP-NER), which uses frozen pre-trained language models and domain-specific prefixes to transfer knowledge from multiple source domains to low-resource target domains, achieving better performance on the Cross-NER benchmark compared to previous methods.
Cross-domain NER is a challenging task to address the low-resource problem in practical scenarios. Previous typical solutions mainly obtain a NER model by pre-trained language models (PLMs) with data from a rich-resource domain and adapt it to the target domain. Owing to the mismatch issue among entity types in different domains, previous approaches normally tune all parameters of PLMs, ending up with an entirely new NER model for each domain. Moreover, current models only focus on leveraging knowledge in one general source domain while failing to successfully transfer knowledge from multiple sources to the target. To address these issues, we introduce Collaborative Domain-Prefix Tuning for cross-domain NER (CP-NER) based on text-to-text generative PLMs. Specifically, we present text-to-text generation grounding domain-related instructors to transfer knowledge to new domain NER tasks without structural modifications. We utilize frozen PLMs and conduct collaborative domain-prefix tuning to stimulate the potential of PLMs to handle NER tasks across various domains. Experimental results on the Cross-NER benchmark show that the proposed approach has flexible transfer ability and performs better on both one-source and multiple-source cross-domain NER tasks. Codes are available in https://github.com/zjunlp/DeepKE/tree/main/example/ner/cross.