AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Finding Groups of Cross-Correlated Features in Bi-View Data

Miheer Dewaskar, John Palowitch, Mark He, Michael I. Love, Andrew B. Nobel

JMLR 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
account for intra-correlations among features of the same type when testing (BSP rather than CONDOR/Matrix-eQTL)BSP · Simulation (n=200)average edge-error: 0.09 -> 0.041Improved
increase sample size from 200 to 600BSP · SimulationJaccard recovery of target bimodules: 8% -> 25% of BSP bimodules overlap two or more target bimodulesHurt
increase sample size from 200 to 600BSP · Simulationrecall of weaker bimodules: recovers bimodules with cross-correlation strength above 0.3 (previously 0.4)Improved
increase sCCA penalty lambda from 0.01 to 0.04sCCA · Simulation (n=200)average edge-error: 0.24 -> 0.89Hurt
increase sCCA penalty lambda from 0.01 to 0.02sCCA · Simulation (n=200)99th percentile recall of target bimodules: 0.1 -> 0.98Improved
initialize search from 10% of SNPs rather than half (ablation)BSP · GTEx thyroid eQTLeffective number of bimodules: 3304 -> 3258No clear effect
permute sample labels of expression and covariatesBSP · GTEx thyroid eQTLnumber of bimodules found: 3344 -> 5-12Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy