A survey of biomedical relation extraction systems: techniques, domains, and benchmarks

Biomedical literature experiences exponential growth, making the automatic extraction of relations between entities crucial for diverse applications. Understanding the interplay of genes, diseases, and drugs through gene-disease, drug-disease, and drug-protein interaction identification is paramount. This thesis investigates the landscape of BioRE technologies, systematically analyzing systems across these key relation categories. The study delves into the features of each surveyed system, encompassing their structures, training methodologies (supervised, semi-supervised, unsupervised), and performance benchmarks on established datasets. Additionally, we compiled a valuable resource list of publicly available datasets and essential biomedical text processing tools. By comparing the strengths and weaknesses of different approaches, this survey aims to provide a comprehensive overview of current advancements and future directions in BioRE, fostering a more profound understanding of this pivotal field within biomedical text mining.