HyperPPO: A scalable method for finding small policies for robotic control
HyperPPO is a reinforcement learning algorithm that uses graph hypernetworks to estimate the weights of multiple neural architectures at once. It is designed to find smaller, more efficient neural network…
Continue reading