PANER: A Paraphrase-Augmented Framework for Low-Resource Named Entity Recognition

2025-10-20

AI Summary

In low-resource named entity recognition (NER), scarce annotated data and poor generalization of zero-shot/instruction-tuning methods to domain-specific entities pose significant challenges. To address these, this paper proposes a lightweight few-shot NER framework. Its core contributions are: (1) an instruction-tuning template with a simplified output format, designed to maximize the utility of large language models’ context windows; and (2) a context-aware sentence paraphrasing strategy that preserves entity semantics for data augmentation, substantially improving few-shot generalization. The method integrates instruction tuning, contextual paraphrasing, few-shot learning, and zero-shot transfer. Evaluated on the CrossNER benchmark, it achieves an average few-shot F1 score of 80.1%, outperforming the strongest baseline by 17.0 percentage points and establishing new state-of-the-art performance among low-resource NER approaches.

Technology Category

• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Common-Sense Reasoning

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Named Entity Recognition (NER) is a critical task that requires substantial annotated data, making it challenging in low-resource scenarios where label acquisition is expensive. While zero-shot and instruction-tuned approaches have made progress, they often fail to generalize to domain-specific entities and do not effectively utilize limited available data. We present a lightweight few-shot NER framework that addresses these challenges through two key innovations: (1) a new instruction tuning template with a simplified output format that combines principles from prior IT approaches to leverage the large context window of recent state-of-the-art LLMs; (2) introducing a strategic data augmentation technique that preserves entity information while paraphrasing the surrounding context, thereby expanding our training data without compromising semantic relationships. Experiments on benchmark datasets show that our method achieves performance comparable to state-of-the-art models on few-shot and zero-shot tasks, with our few-shot approach attaining an average F1 score of 80.1 on the CrossNER datasets. Models trained with our paraphrasing approach show consistent improvements in F1 scores of up to 17 points over baseline versions, offering a promising solution for groups with limited NER training data and compute power.

Problem

• Addressing low-resource NER with limited annotated data
• Improving generalization to domain-specific entities in NER
• Enhancing NER performance through strategic data augmentation

Innovation

• Lightweight few-shot NER framework with simplified output format
• Instruction tuning template leveraging large context window LLMs
• Strategic paraphrasing data augmentation preserving entity information