Incorporate outlier detection capabilities into AdapDISCOM

Develop an outlier detection extension of AdapDISCOM that augments its multimodal covariance-based sparse regression framework to identify and handle outliers in addition to addressing block-wise missingness and measurement error.

Background

AdapDISCOM focuses on mitigating block-wise missingness and measurement error, with robust variants for heavy-tailed distributions. However, explicit outlier detection is not part of the current framework.

The authors point to enhancing AdapDISCOM with outlier detection as an open direction, motivated by related robust modeling approaches.