AdaDeep: A usage-driven, automated deep model compression framework for enabling ubiquitous intelligent mobiles

• Main Authors: LIU, Sicong, DU, Junzhao, NAN, Kaiming, ZHOU, Zimu, LIU, Hui, WANG, Zhangyang, LIN, Yingyan
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2021
• Subjects: Optimization; Mobile Computing; Mobile Handsets; Mobile Applications; Energy Storage; Measurement; Computational Modeling; Artificial Intelligence and Robotics; Software Engineering
• Online Access: https://ink.library.smu.edu.sg/sis_research/6648; https://ink.library.smu.edu.sg/context/sis_research/article/7651/viewcontent/tmc21_liu.pdf
• Institution: Singapore Management University

Recent breakthroughs in deep neural networks (DNNs) have fueled a tremendously growing demand for bringing DNN-powered intelligence into mobile platforms. While the potential of deploying DNNs on resource-constrained platforms has been demonstrated by DNN compression techniques, the current practice suffers from two limitations: 1) merely stand-alone compression schemes are investigated even though each compression technique only suit for certain types of DNN layers; and 2) mostly compression techniques are optimized for DNNs’ inference accuracy, without explicitly considering other application-driven system performance (e.g., latency and energy cost) and the varying resource availability across platforms (e.g., storage and processing capability). To this end, we propose AdaDeep, a usage-driven, automated DNN compression framework for systematically exploring the desired trade-off between performance and resource constraints, from a holistic system level. Specifically, in a layer-wise manner, AdaDeep automatically selects the most suitable combination of compression techniques and the corresponding compression hyperparameters for a given DNN. Thorough evaluations on six datasets and across twelve devices demonstrate that AdaDeep can achieve up to 18.6×18.6×18.6× latency reduction, 9.8×9.8×9.8× energy-efficiency improvement, and 37.3×37.3×37.3× storage reduction in DNNs while incurring negligible accuracy loss. Furthermore, AdaDeep also uncovers multiple novel combinations of compression techniques.