An active-learning method using a bias-variance decomposition and a 'cobias-covariance' matrix with eigendecomposition-based batch selection outperforms BALD and least confidence on synthetic noise benchmarks.
sctrends: A living review of commercial single-cell and spatial’omic technologies
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
When three experiments are better than two: Avoiding intractable correlated aleatoric uncertainty by leveraging a novel bias--variance tradeoff
An active-learning method using a bias-variance decomposition and a 'cobias-covariance' matrix with eigendecomposition-based batch selection outperforms BALD and least confidence on synthetic noise benchmarks.