A key challenge for scaling our approach to different domains is its dependency on high-quality external ontologies and knowledge bases. This factor limits the scope of our analyses across biomedical domains.
More efficient methods for populating the natural language syllogistic arguments could be investigated in future work, involving automated NLP methods, such as those used in RepoDB, MSI, Hetionet, DrugMechDB, and INDRA.
However, these approaches still face challenges in balancing precision and generalization, particularly for complex reasoning tasks in biomedicine. Further improvements are necessary to develop scalable resources and more adaptable NLP techniques for real-world applications.
Collection
[
|
...
]