When Covariance Is Not Enough: Kernel-Based Regularization for Single-Cell Counterfactuals
arXiv:2611.13748v1 Announce Type: cross Abstract: Counterfactual editing of single cells, predicting how an observed cell would look had one factor been different, enables in-silico screening where experiments are infeasible, but is useful only if the edit is faithful to the individual and not merely to the condition.