고종환 교수

Convolutional neural networks (CNNs) require high computation and memory demand for training. This paper presents the design of a frequency-domain accelerator for energy-efficient CNN training. With Fourier representations of parameters, we replace convolutions with simpler pointwise multiplications. To eliminate the Fourier transforms at every layer, we train the network entirely in the frequency domain using approximate frequency-domain nonlinear operations. We further reduce computation and m

Neural networks generally require significant memory capacity/bandwidth to store/access a large number of synaptic weights. This paper presents an application of JPEG image encoding to compress the weights by exploiting the spatial locality and smoothness of the weight matrix. To minimize the loss of accuracy due to JPEG encoding, we propose to adaptively control the quantization factor of the JPEG algorithm depending on the error-sensitivity (gradient) of each weight. With the adaptive compress

In a wireless video sensor system, encoder and transmitter parameters should be jointly adjusted for the optimal tradeoff between video quality and energy consumption under variable channel conditions. The optimization of the system can be further enhanced by exploiting relative importance of moving objects in remote surveillance applications. This paper presents an energy-efficient wireless video sensor node for remote surveillance using content-aware pre-processing and an energy- and content-a

Abstract Reversible metal‐filamentary mechanism has been widely investigated to design an analog resistive switching memory (RSM) for neuromorphic hardware‐implementation. However, uncontrollable filament‐formation, inducing its reliability issues, has been a fundamental challenge. Here, an analog RSM with 3D ion transport channels that can provide unprecedentedly high reliability and robustness is demonstrated. This architecture is realized by a laser‐assisted photo‐thermochemical process, comp

This paper presents a single-chip image sensor node with energy harvesting from the pixel array. The design includes a 128 × 96 pixel array that can be reconfigured to form an on-chip photovoltaic cell to harvest energy. An onchip power management unit harvests energy from the pixel array, and delivers multiple regulated output voltage domains to power the sensor, image processor, and memory. The image processor is a low-overhead moving object detection unit to reduce the volume of transmitted d

Abstract Continued research on the epidermal electronic sensor aims to develop sophisticated platforms that reproduce key multimodal responses in human skin, with the ability to sense various external stimuli, such as pressure, shear, torsion, and touch. The development of such applications utilizes algorithmic interpretations to analyze the complex stimulus shape, magnitude, and various moduli of the epidermis, requiring multiple complex equations for the attached sensor. In this experiment, we

Neural networks generally require significant memory capacity/bandwidth to store/access a large number of synaptic weights. This paper presents design of an energy-efficient neural network inference engine based on adaptive weight compression using a JPEG image encoding algorithm. To maximize compression ratio with minimum accuracy loss, the quality factor of the JPEG encoder is adaptively controlled depending on the accuracy impact of each block. With 1% accuracy loss, the proposed approach ach

This paper introduces partitioning an inference task of a deep neural network between an edge and a host platform in the IoT environment. We present a DNN as an encoding pipeline, and propose to transmit the output feature space of an intermediate layer to the host. Encoding of the feature space is proposed to enhance the maximum input rate supported by the edge platform and/or reduce the energy of the edge platform. Simulation results show that partitioning a DNN coupled with feature space enco

While deep neural networks have shown powerful performance in many audio applications, their large computation and memory demand has been a challenge for real-time processing. In this paper, we study the impact of scaling the precision of neural networks on the performance of two common audio processing tasks, namely, voice-activity detection and single-channel speech enhancement. We determine the optimal pair of weight/neuron bit precision by exploring its impact on both the performance and pro

Fast and robust voice-activity detection is critical to efficiently process speech. While deep-learning based methods to detect voice have shown competitive accuracies, the best models in the literature incur over a 100 ms latency on commodity processors. Such delays are unacceptable for real-time speech processing. In this paper, we study the impact of lowering the representation precision of the neural-network weights and neurons on both the accuracy and delay of voice-activity detection. Base

Wireless image sensor nodes are required to deliver better visual information of the region-of-interest (ROI) under tight energy constraints. The energy-quality scalability of the sensor node can be improved by incorporating ROI-based image processing. This paper presents an energy-quality scalable wireless image sensor node using ROI-based processing for object-based surveillance. After detecting the ROI using a low-power noise-robust method, frame images are encoded by a low-complexity ROI cod