Towards the Efficient Multi-Platform Execution of Deep Neural Networks | IEEE Conference Publication | IEEE Xplore