selected publications conference paper Efficient Hardware Approximation for Bit-Decomposition Based Deep Neural Network Accelerators. . 2022 Enhancing the Utilization of Dot-Product Engines in Deep Learning Accelerators. . 2020