AGV: Evolve a Deep Network to drive an autonomous Vehicle involving end-to-end Learning
This topic can be conducted both in English and Hungarian ! / A téma magyarul és angolul is felvehető !
The aim of the project is to develop an end-to-end learning model that controls an autonomous vehicle. By end-to-end learning we mean that the sensors and the motors of the vehicle are directly connected to the neural network. The evolution of the network is an iterative process and starts with a population of networks. In each round, a new population is generated from the best performing models. The goal of this iteration is to find the best performing network that is able to drive the vehicle.
For more information, see examples of deep learning combined with genetic algorithm: