Toward Energy–Quality Scaling in Deep Neural Networks
Jeff Anderson, Yousra Alkabani, Tarek El‐Ghazawi · IEEE Design and Test · 2019
Editor's notes: This article surveys the latest advances in neural network (NN) architectures by applying them to the task of energy-quality scaling. Results show that, while coarse scaling is possible with existing NN architectures, fine-grain scaling is needed for fog computing efforts, and further work should focus on hybrid NN architecture development.