Putting the “Machine” Back in Machine Learning: The Case for Hardware-ML Model Co-design

NeurIPS 2019