A Structure-aware Generative Model for Biomedical Event Extraction

2024-08-13

arXiv.org

AI Summary

Biomedical event extraction (BEE) faces significant challenges in modeling nested and overlapping events—constituting over 20% of annotations—while existing generative models neglect inter-instance structural dependencies. To address this, we propose a structure-aware generative framework. First, we explicitly model cross-instance event structural relations within the generative paradigm—a novel departure from prior work. Second, we design an event-structure prompting mechanism that jointly encodes label semantics and argument dependency topology; further, we leverage large language model knowledge distillation to construct structured prefixes that enhance generation fidelity. Our approach achieves state-of-the-art performance on MLEE and GE11, matches the best discriminative models on PHEE, and yields substantial gains in nested and overlapping event identification accuracy.

Technology Category

• Machine Learning: Structured Learning
• Natural Language Processing: Generation
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Bridging structured and unstructured data

Abstract

Biomedical Event Extraction (BEE) is a challenging task that involves modeling complex relationships between fine-grained entities in biomedical text. BEE has traditionally been formulated as a classification problem. With recent advancements in large language models (LLMs), generation-based models that cast event extraction as a sequence generation problem have attracted attention in the NLP research community. However, current generative models often overlook cross-instance information in complex event structures, such as nested and overlapping events, which constitute over 20% of events in benchmark datasets. In this paper, we propose GenBEE, an event structure-aware generative model that captures complex event structures in biomedical text for biomedical event extraction. GenBEE constructs event prompts that distill knowledge from LLMs to incorporate both label semantics and argument dependency relationships. In addition, GenBEE generates prefixes with event structural prompts to incorporate structural features to improve the model's overall performance. We have evaluated the proposed GenBEE model on three widely used BEE benchmark datasets, namely MLEE, GE11, and PHEE. Experimental results show that GenBEE has achieved state-of-the-art performance on the MLEE and GE11 datasets, and achieved competitive results when compared to the state-of-the-art classification-based models on the PHEE dataset.

Problem

• Improves biomedical event extraction accuracy
• Addresses nested and overlapping event structures
• Leverages large language models for sequence generation

Innovation

• Generative model captures complex event structures
• Event prompts distill knowledge from LLMs
• Prefixes with structural prompts enhance performance