STAYKATE: Hybrid In-Context Example Selection Combining Representativeness Sampling and Retrieval-based Approach -- A Case Study on Science Domains

TL;DR

STAYKATE is a hybrid in-context example selection method that combines representativeness sampling and retrieval to improve scientific information extraction with large language models, especially for challenging entity types.

Contribution

The paper introduces STAYKATE, a novel hybrid selection approach that outperforms traditional methods in domain-specific scientific information extraction tasks.

Findings

STAYKATE outperforms traditional supervised and existing selection methods.

Performance improvements are significant for difficult entity types.

Effective across three domain-specific datasets.

Abstract

Large language models (LLMs) demonstrate the ability to learn in-context, offering a potential solution for scientific information extraction, which often contends with challenges such as insufficient training data and the high cost of annotation processes. Given that the selection of in-context examples can significantly impact performance, it is crucial to design a proper method to sample the efficient ones. In this paper, we propose STAYKATE, a static-dynamic hybrid selection method that combines the principles of representativeness sampling from active learning with the prevalent retrieval-based approach. The results across three domain-specific datasets indicate that STAYKATE outperforms both the traditional supervised methods and existing selection methods. The enhancement in performance is particularly pronounced for entity types that other methods pose challenges.