1 paper
Nuthan Mummani, Simran Ketha, Venkatakrishnan Ramaswamy
In the supervised classification setting, during inference, deep networks typically make multiple predictions. For a pair of such predictions (that are in the top-k predictions), t…