Abstract:Few-shot Relation Extraction (FSRE) aims to extract relational facts from a sparse set of labeled corpora. Recent studies have shown promising results in FSRE by employing Pre-trained Language Models (PLMs) within the framework of supervised contrastive learning, which considers both instances and label facts. However, how to effectively harness massive instance-label pairs to encompass the learned representation with semantic richness in this learning paradigm is not fully explored. To address this gap, we introduce a novel synergistic anchored contrastive pre-training framework. This framework is motivated by the insight that the diverse viewpoints conveyed through instance-label pairs capture incomplete yet complementary intrinsic textual semantics. Specifically, our framework involves a symmetrical contrastive objective that encompasses both sentence-anchored and label-anchored contrastive losses. By combining these two losses, the model establishes a robust and uniform representation space. This space effectively captures the reciprocal alignment of feature distributions among instances and relational facts, simultaneously enhancing the maximization of mutual information across diverse perspectives within the same relation. Experimental results demonstrate that our framework achieves significant performance enhancements compared to baseline models in downstream FSRE tasks. Furthermore, our approach exhibits superior adaptability to handle the challenges of domain shift and zero-shot relation extraction. Our code is available online at <a class="link-external link-https" href="https://github.com/AONE-NLP/FSRE-SaCon" rel="external noopener nofollow">this https URL</a>.

Preserving Generalization of Language models in Few-shot Continual Relation Extraction

Consistent Representation Learning for Continual Relation Extraction

Less is More: Rethinking State-of-the-art Continual Relation Extraction Models with a Frustratingly Easy but Effective Approach

Empowering Few-Shot Relation Extraction with The Integration of Traditional RE Methods and Large Language Models

Joint data augmentation and knowledge distillation for few-shot continual relation extraction

Rationale-Enhanced Language Models are Better Continual Relation Learners

Few-Shot Relation Extraction With Automatically Generated Prompts

Refining Sample Embeddings with Relation Prototypes to Enhance Continual Relation Extraction

Envisioning Class Entity Reasoning by Large Language Models for Few-shot Learning

Document-Level In-Context Few-Shot Relation Extraction via Pre-Trained Language Models

Enhancing Continual Relation Extraction Via Classifier Decomposition

Learning Discriminative Semantic and Multi-view Context for Domain Adaptive Few-Shot Relation Extraction.

Learning Robust Representations for Continual Relation Extraction Via Adversarial Class Augmentation

Synergistic Anchored Contrastive Pre-training for Few-Shot Relation Extraction

Defying Forgetting in Continual Relation Extraction Via Batch Spectral Norm Regularization

DP-CRE: Continual Relation Extraction via Decoupled Contrastive Learning and Memory Structure Preservation

How to Unleash the Power of Large Language Models for Few-shot Relation Extraction?

RAPL: A Relation-Aware Prototype Learning Approach for Few-Shot Document-Level Relation Extraction

Grasping the Essentials: Tailoring Large Language Models for Zero-Shot Relation Extraction

Serial Contrastive Knowledge Distillation for Continual Few-shot Relation Extraction

Improving Recall of Large Language Models: A Model Collaboration Approach for Relational Triple Extraction