No Prior, No Leakage: Revisiting Reconstruction Attacks in Trained Neural Networks

MIACVAAMLRe-assign community