Adaptive Symmetrization of the KL Divergence An approach to non-adversarially minimize the symmetric Jeffreys divergence. Collective Action for Fairness Can people collaborate to make ML fair?