JojoSCL combines instance-level and cluster-level contrastive learning with a SURE-based shrinkage loss for scRNA-seq clustering and reports improved ARI and NMI on most of ten benchmark datasets.
Deep learning,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
JojoSCL: Shrinkage Contrastive Learning for single-cell RNA sequence Clustering
JojoSCL combines instance-level and cluster-level contrastive learning with a SURE-based shrinkage loss for scRNA-seq clustering and reports improved ARI and NMI on most of ten benchmark datasets.