Revista de teledetección y SIG

Abstracto

Deep learning for tiny devices

Anudeepsekhar Bolimera

The talk aims to provide a timely review of the fast-paced field of deep learning and artificial intelligence technologies for mobile devices. Enabling deep learning on mobile devices provides several advantages such as data privacy and quicker response times. While traditional computation paradigms rely mostly on cloud computing and connectivity to the cloud, recent breakthroughs in the field have enabled numerous mobile applications. Mobile devices are constrained by the size, weight, area and power considerations and also their computational capabilities. Addressing certain key challenges in deploying deep learning to mobile devices author aims to present the current state of the art techniques and algorithms in their relation to algorithm optimizations that simplify computation while retaining performance accuracy. They also aim to briefly present various applications of these algorithms in industries, ranging from robotics and healthcare to autonomous driving and defense supporting them with implementations and benchmark.