Work

An Approximate Memory Architecture for a Reduction of Refresh Power Consumption in Deep Learning Applications