AdAM: Adaptive Fault-Tolerant Approximate Multiplier for Edge DNN Accelerators

The remainder of the paper is organized as follows. Section The role of Deep Neural Networks (DNNs) in a wide range II summarizes related works, the proposed method is presented of safety-and mission-critical applications (e.g., autonomous in Section III, Section IV provides the experimental setup and driving) is expanding. Therefore, deployment of a DNN accelerator discusses the results, and finally, the work is concluded in requires addressing the trade-off between different Section V. design parameters and reliability [1] [2].