Mutual information (MI) quantifies the dependence between features and target variables and is widely used in feature selection for downstream tasks. However, what works globally often fails locally and different feature subsets may best fit different subgroups. The paper “MISFEAT: Feature Selection for Subgroups with Mutual Information Estimation“, co-authored by Bar Genossar, Thinh On, Md Mouinul Islam, Ben Eliav, Senjuti Basu Roy and Avigdor Gal, introduces a framework that models feature-subgroup-target interactions as a multiplex graph and applies a heterogeneous graph neural network to propagate information between feature combinations both within and across subgroups to efficiently identify top-K predictive feature subsets per subgroup while addressing key scalability challenges. Join us in Montreal, where the paper will be presented as part of ICDE, the International Data Engineering Conference.