AI And Energy Efficiency

Remarkable progress has been reported in the deployment of artificial intelligence and machine learning applications in a broad range of capabilities such as healthcare, game playing, image recognition, and machine translation. At the same time, the growing usage of AI systems demand for more energy contributed to increasing CO2 emissions. Most of the studies are focusing on increased accuracy rather than energy efficiency of these models. In this research, we are investigating the energy cost associated with AI models, and techniques to help increase the energy efficiency of AI models, and conduct empirical experiments to validate the goodness of the techniques identified. The proposed techniques will help the AI developers, engineers, and community achieve increased energy efficiency of AI models. In addition, we also identify an approach that will allow the software development community to create an energy-efficient configuration of software through the use of machine learning