Federated kMeans, PCA+kMeans, and LSH methods can approximate centralized clustering for molecular data diversity when evaluated with chemistry-informed metrics like the introduced SF-ICF.
Deep clustering of small molecules at large-scale via variational autoencoder embedding and k-means
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Insights into the Unknown: Federated Data Diversity Analysis on Molecular Data
Federated kMeans, PCA+kMeans, and LSH methods can approximate centralized clustering for molecular data diversity when evaluated with chemistry-informed metrics like the introduced SF-ICF.