Dataset for N-ary Relation Extraction of Drug Combinations

Abstract

Combination therapies have become the standard of care for diseases such as cancer, tuberculosis, malaria and HIV. However, the combinatorial set of available multi-drug treatments creates a challenge, particularly in the presence of antagonistic drug combinations that may lead to negative patient outcomes. To assist medical professionals in identifying beneficial drug-combinations, we construct an expert-annotated dataset for extracting information about the efficacy of drug combinations from the scientific literature. Beyond its practical utility, the dataset also presents a unique NLP challenge, as it is the first relation extraction dataset consisting of variable-length relations. Furthermore, the relations in this dataset predominantly require language understanding beyond the sentence level, adding to the challenge of this task. We provide a strong baseline model and identify clear areas for further improvement. We release our dataset and code (https://anonymous.4open.science/r/drug-synergy-models--C8B7/README.md) publicly to encourage the NLP community to participate in this task.

Tasks

Reproductions