Reducing power consumption is crucial for edge devices using convolutional neural network (CNN). The zero-skipping approach for CNNs is a processing technique widely known for its relatively low power consumption and high speed. This approach stops multiplication and accumulation (MAC) when the multiplication results of the input data and weight are zero. However, this technique requires large logic circuits with around 5% overhead, and the average rate of MAC stopping is approximately 30%. In this paper, we propose a precise zero-skipping method that uses input data and simple logic circuits to stop multipliers and accumulators precisely. We also propose an active data-skipping method to further reduce power consumption by slightly degrading recognition accuracy. In this method, each multiplier and accumulator are stopped by using small values (e.g., 1, 2) as input. We implemented single shot multi-box detector 500 (SSD500) network model on a Xilinx ZU9 and applied our proposed techniques. We verified that operations were stopped at a rate of 49.1%, recognition accuracy was degraded by 0.29%, power consumption was reduced from 9.2 to 4.4 W (-52.3%), and circuit overhead was reduced from 5.1 to 2.7% (-45.9%). The proposed techniques were determined to be effective for lowering the power consumption of CNN-based edge devices such as FPGA.

We propose new electro-optic modulators using a double antenna-coupled electrode structure for radio-over-fiber systems. The proposed modulators are composed of a pair of patch antennas and a standing-wave resonant electrode. By utilizing a pair of patch antennas on SiO2 substrates and a coupled-microstrip line resonant electrode on a LiNbO3 substrate with a symmetric configuration, high-efficiency optical modulation is obtainable for 24 optical waveguides at the same time. The proposed modulators were designed at 58 GHz and their basic operations were demonstrated successfully with an improvement of 9 dB compared to a single antenna-coupled electrode device on a LiNbO3 substrate in our previous work.