Access to Data or Embedding #1
Description
Hello,
Thank you for your on this project. I am particularly interested in the methods and data you've used in your experiments.
- Access to Data/Embeddings: Is it possible to obtain access to the embeddings or the dataset utilized in your experiments? Access to these resources would greatly assist others in the community by enabling replication and further research based on your findings.
- Could you clarify whether the SAPBert model is solely used for generating entity mentions (names only) and concept mentions from vocabularies like SNOMED and ICD-10? It appears that the system relies on SAPBert as a retriever. If so, could you explain how this setup is distinct from using SAPBert semantic search as mentioned in the SAPBert evaluation code on GitHub? Specifically, I am interested in understanding the comparisons drawn in your results.