1 paper
Aiden Durrant, Rahul Baburajan, Georgios Leontidis
While recent methods like that of the Poincaré ResNet have demonstrated the ability to learning visual representations directly in hyperbolic space, their optimisation remains a ch…