Abstract
Smart computing devices continue to increase in popularity, availability, and functionality. A smart computing device, such as a security camera, collects or senses information and processes that information. In processing the information, the smart computing device engages in machine-learning processes to make inferences or predictions based on the collected information. Hardware components, energy resources, and even the environment in which the smart computing device resides affect the capability of the smart computing device to process the collected information and make the inferences or predictions. Tracking and optimizing energy consumption can increase the efficiency and efficacy of smart computing devices. Energy consumption can be optimized by dynamically adjusting the sensing protocols or information collection capabilities of the smart computing device or by dynamically adjusting the machine-learning processes executed by the smart computing device.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Nagarajan, Gurunathan, "Using Energy Consumption Metrics to Make Machine-Learning Models Energy Optimal", Technical Disclosure Commons, (July 20, 2018)
https://www.tdcommons.org/dpubs_series/1349