https://iris-lab.skku.edu/project/ Projects Wowchemy (https://wowchemy.com)en-usTue, 22 Dec 2020 22:12:44 +0900 https://iris-lab.skku.edu/media/icon_hub30d19cd1ec39a1dd8652017ce0a280f_82173_512x512_fill_lanczos_center_3.png https://iris-lab.skku.edu/project/ https://iris-lab.skku.edu/project/ml_mem/ Tue, 22 Dec 2020 22:12:44 +0900 https://iris-lab.skku.edu/project/ml_mem/ <p>Our research focuses on memory-based deep learning techniques, including SRAM/ReRAM-based deep learning processing-in-memory (PIM) and Flash memory based robust deep learning inference techniques.</p> https://iris-lab.skku.edu/project/mm_3d/ Sun, 22 Nov 2020 22:12:44 +0900 https://iris-lab.skku.edu/project/mm_3d/ <p>As three-dimensional (3D) vision data can provide abundant spatial information, it is being widely used in many areas, including autonomous driving and mobile robots. We aim to develop efficient and accurate 3D data processing techniques for classification, segmentation, pose estimation, etc.</p> https://iris-lab.skku.edu/project/hw_mlsensor/ Sun, 22 Dec 2019 22:12:44 +0900 https://iris-lab.skku.edu/project/hw_mlsensor/ <p>To maximize the performance and efficiency of deep learning based data processing, we are currently exploring sensor platform design optimized for deep learning by leveraging the interactions between a sensor platform and a DNN.</p> https://iris-lab.skku.edu/project/ml_acceleration/ Sat, 22 Dec 2018 22:12:44 +0900 https://iris-lab.skku.edu/project/ml_acceleration/ <p>Deep neural networks (DNNs) are widely adopted at the IoT edge devices to enable more intelligence. However, the biggest challenge is their storage demand and computational complexity. As DNNs contain a large number of synaptic weights, the memory demand is a key challenge for application of DNNs, especially for memory-constrained platforms such as mobile systems. Another bottleneck of DNNs is the computation demand, mostly due to their large number of multiplications in convolution layers. Therefore, reducing the storage and computation demand of neural networks is critical, specifically, to support in-field and on-chip training and inference. To enable deep learning based intelligent multimedia processing at the IoT edge devices with limited hardware resource, the research targets for neural network design with lower storage and computation demand.</p> https://iris-lab.skku.edu/project/hw_digital/ Sat, 22 Dec 2018 22:12:44 +0900 https://iris-lab.skku.edu/project/hw_digital/ <p>Tolerating timing error due to power supply noise (PSN) in digital circuits can be done with adding voltage margins. We designed guidelines to avoid overdesign due to PSN especially for the low-cost IoT devices. Specifically, we presented an accurate time-domain behavioral model of timing slack variation due to PSN accounting for the clock-data compensation.</p> https://iris-lab.skku.edu/project/ml_collaboration/ Sat, 22 Dec 2018 22:12:44 +0900 https://iris-lab.skku.edu/project/ml_collaboration/ <p>Based on the trade-off study between energy, accuracy, and throughput of an edge platform with an embedded DNN, we are currently exploring collaborative DNN inference between the edge and the host to optimally utilize the available bandwidth and energy.</p> https://iris-lab.skku.edu/project/hw_security/ Sat, 22 Dec 2018 22:12:44 +0900 https://iris-lab.skku.edu/project/hw_security/ <p>The physical security is a key challenge for the resource-constrained edge platforms. A key challenge is to enable secure as well as ultra-low-power hardware. The research seeks to understand the interactions between low-power and security in edge devices, and explore innovations to enhance security at minimal power cost. Specifically, the research investigated the interactions between design of area-/power- efficient crypto engines and their side-channel resistance. The goal is to enable side-channel secure encryption in resource-constrained platforms such as sensors.</p> https://iris-lab.skku.edu/project/hw_sensor/ Sat, 22 Dec 2018 22:12:44 +0900 https://iris-lab.skku.edu/project/hw_sensor/ <p>The objective of this work is to design a self-powered, environment-adaptive sensor node that maintains a target Quality-of-Service (QoS) in a time-varying environment. A wireless image sensor node will be designed that incorporates a CMOS imager, digital signal processing unit, and RF transreceiver and is powered using energy harvested from the environment. The circuit innovations for the individual components will be coupled with system design and on-line real-time control principles to approach this highly challenging goal of developing a self-powered image sensor node. The self-powered sensor node and reliable energy-efficient image transmission principles created in this work will allow deployment of image sensors and communication networks to cyber-physical systems in various military as well as civilian applications.</p> https://iris-lab.skku.edu/project/mm_audio/ Sat, 22 Dec 2018 22:12:44 +0900 https://iris-lab.skku.edu/project/mm_audio/ <p>Our research focuses on front-end audio processing techniques for speech processing, such as voice activity detection (VAD), noise suppression, direction of arrival (DoA) estimation, and classification.</p> https://iris-lab.skku.edu/project/mm_image/ Sat, 22 Dec 2018 22:12:44 +0900 https://iris-lab.skku.edu/project/mm_image/ <p>Machine learning and deep learning have made rapid progress in many computer vision applications over a short period. We focus on diverse techniques for image and video processing powered by machine/deep learning.</p> https://iris-lab.skku.edu/project/ml_robust/ Sat, 22 Dec 2018 22:12:44 +0900 https://iris-lab.skku.edu/project/ml_robust/ <p>One of the challenges of deploying deep neural networks in sensor platforms is the variations in the input images; structural noise (adversarial images), inherent random noise (image perturbation), input scene variability, and weight value errors. To enhance the robustness of deep learning inference to the variability, we are designing variation-robust deep learning techniques.</p> https://iris-lab.skku.edu/project/mm_lowpower/ Fri, 22 Dec 2017 22:12:44 +0900 https://iris-lab.skku.edu/project/mm_lowpower/ <p>A critical goal in the image sensor node design is to deliver high-quality visual information under stringent energy and bandwidth constraints. This goal becomes more challenging under dynamic conditions such as environmental noise and variations in a wireless channel condition. We tackle this challenge by designing low-power energy-quality scalable image processing techniques.</p>