A Hessian-informed hyperparameter optimization for differential learning rate

Re-assign community