Skip to main content
. Author manuscript; available in PMC: 2024 Apr 24.
Published in final edited form as: J Biomed Inform. 2022 Mar 26;129:104059. doi: 10.1016/j.jbi.2022.104059

Fig. 3.

Fig. 3.

Architecture of the re-ranking model. Pairs of document and candidate HPO terms are inputted into the model. The text in the document and HPO term (or term definition) are merged into the BlueBERT model, while the HPO IDs are fed into the selection-embedding module. The outputs of these two streams are concatenated before the linear transformation module.