ABSTRACT
The chapter introduces device constrained modeling framework, the purpose of which is to enable engineering designs to holistically look three major engineering characteristics so as to deploy Machine Learning models in extreme constrained environments. The design characteristics include: AI/ML Model Accuracy, Connectivity and Hardware. Next, the chapter introduces tradeoff modeling between three major hardware design characteristics to fine tune the necessary adjustments in design characteristic to compensate any tradeoff in function or hardware features. And finally, the chapter concludes by introducing factors such as environment & perturbation of data in crossing the Chasm to democratize the AI.
