Learning from Negative Samples in Biomedical Generative Entity Linking

TL;DR

This paper introduces ANGEL, a novel framework for biomedical entity linking that trains generative models using negative samples, leading to improved accuracy by explicitly learning from hard negatives.

Contribution

ANGEL is the first framework to incorporate negative samples into training for biomedical generative entity linking, enhancing model performance.

Findings

Models outperform previous baselines by up to 1.4% top-1 accuracy.

Performance improves further by 1.7% when integrated into pre-training.

Effective in both pre-training and fine-tuning stages.

Abstract

Generative models have become widely used in biomedical entity linking (BioEL) due to their excellent performance and efficient memory usage. However, these models are usually trained only with positive samples, i.e., entities that match the input mention's identifier, and do not explicitly learn from hard negative samples, which are entities that look similar but have different meanings. To address this limitation, we introduce ANGEL (Learning from Negative Samples in Biomedical Generative Entity Linking), the first framework that trains generative BioEL models using negative samples. Specifically, a generative model is initially trained to generate positive entity names from the knowledge base for given input entities. Subsequently, both correct and incorrect outputs are gathered from the model's top-k predictions. Finally, the model is updated to prioritize the correct predictions through preference optimization. Our models outperform the previous best baseline models by up to an average top-1 accuracy of 1.4% on five benchmarks. When incorporating our framework into pre-training, the performance improvement increases further to 1.7%, demonstrating its effectiveness in both the pre-training and fine-tuning stages. The code and model weights are available at https://github.com/dmis-lab/ANGEL.