Learning to play Football using Distributional Reinforcement Learning and Depthwise separable convolution feature extraction | IEEE Conference Publication | IEEE Xplore