Abstract:Abstract Motivation Multiple events extraction from biomedical literature is a challenging task for biomedical community. Usually, biomedical event extraction is modeled as two sub-tasks, trigger identification and argument detection. Most existing methods perform these two sub-tasks sequentially, and fail to make full use of the interaction between them, leading to suboptimal results for multiple biomedical events extraction. Results We propose a novel framework of reinforcement learning (RL) for the task of multiple biomedical events extraction. More specifically, trigger identification and argument detection are treated as main-task and subsidiary-task, respectively. Assigning the event type of triggers (in the main-task) is viewed as the action taken in RL, and the result of corresponding argument detection (i.e. the subsidiary-task) for the identified trigger is used for computing the reward of the taken action. Moreover, the result of the subsidiary-task is modeled as part of environment information in RL to help the procedure of trigger identification. In addition, external biomedical knowledge bases are employed for representation learning of biomedical text, which can improve the performance of biomedical event extraction. Results on two widely used biomedical corpora demonstrate that the proposed framework performs better than the selected baselines on the task of multiple events extraction. The ablation test indicates the contributions of RL and external KBs to the performance improvement in the proposed method. In addition, by modeling multiple events extraction under the RL framework, the supervised information is exploited more effectively than the classical supervised learning paradigm. Availability and implementation Source codes will be available at: https://github.com/David-WZhao/BioEE-RL.

Joint Extraction of Biomedical Events Based on Dynamic Path Planning Strategy and Hybrid Neural Network

Conditional Probability Joint Extraction of Nested Biomedical Events: Design of a Unified Extraction Framework Based on Neural Networks

Joint Event Extraction Based on CNN-BiGRU and Attention Mechanism

A Two-Stage Biomedical Event Trigger Detection Method Based on Hybrid Neural Network and Sentence Embeddings

Multi-layer Sequence Labeling-based Joint Biomedical Event Extraction

A neural joint model for entity and relation extraction from biomedical text

Biomedical Event Extraction with Hierarchical Knowledge Graphs

Joint Learning-based Causal Relation Extraction from Biomedical Literature

A tree-like structured perceptron for transition-based biomedical event extraction

Pipelined Biomedical Event Extraction Rivaling Joint Learning

CMBEE: A constraint-based multi-task learning framework for biomedical event extraction

A neural model for joint event detection and prediction

Jointly Events Extraction for Database Alarm Based on Dynamic Matching Strategy and GCN

A tag based joint extraction model for Chinese medical text

A biomedical event extraction method based on fine-grained and attention mechanism

A Novel Joint Framework for Multiple Chinese Events Extraction

BioIE: Biomedical Information Extraction with Multi-head Attention Enhanced Graph Convolutional Network

A system for automatically extracting clinical events with temporal information

An Active Learning Based Hybrid Neural Network for Joint Information Extraction.

Research on Chinese medical named entity recognition based on collaborative cooperation of multiple neural network models

Efficient multiple biomedical events extraction via reinforcement learning