Return to Issue Details Architectural Co-Design and Approximation Strategies for Efficient Deep Neural Network Acceleration in Edge-Oriented Custom Hardware Download Download PDF