1 paper
Vincent Gripon, Matthias Löwe, Franck Vermet
Recently authors have introduced the idea of training discrete weights neural networks using a mix between classical simulated annealing and a replica ansatz known from the statist…