Constraints on the Cloud: why we need machine learning at the Edge – DatacenterDynamics
machine learning machine learning deployment
CCoonnssttrraaiinnttss oonn tthhee CClloouudd:: wwhhyy wwee nneeeedd mmaacchhiinnee lleeaarrnniinngg aatt tthhee EEddggee –– DDaattaacceenntteerrDDyynnaammiiccss
Constraints on the Cloud: why we need machine learning at the Edge  DatacenterDynamicsML has heavy processing demands, but the communications overheads mean we ultimately need it to happen at the very edge: on smartphones.