Computation Efficient Learning Lab.2

Logo
Efficient Machine Learning for Internet of things and edge computing ecosystems
Many applications in the Internet of Things (IoT) and edge computing extract useful information and learn patterns from massive data streams. It poses substantial technical challenges due to the slow response time, scalability issue of learning solutions, and limited resources.
We focus on how to redesign learning by closely modeling the ultimate efficient processor - the human brain. We also explore the potential of learning solutions for future emerging and alternative system designs.
Brain-inspired Hyperdimensional Computing: Congnitive Learning Based on Human Memory Model
Hyperdimensional (HD) computing is an alternative computing method, which processes cognitive tasks in a light-weight and error-torrent way, based on theoretical neuroscience. We work on developing various learning tasks such as classification, regression, clustering, and reinforcement learning using HD computing.
Edge Computing: Learning on Low-power Edge Devices
We build state-of-the-art learning software and hardware on low-power devices. It includes designing self-learning systems capable of autonomous sensing, learning, and actuating on diverse IoT platforms.
Systems for ML and ML for Systems: Alternative Computing
We rethink the role of machine learning (ML) for systems. We explore diverse alternative system solutions such as in-memory computing, near-data computing, and ML-driven system software.


Advisor Professor : Kim, Yeseong
Computation Efficient Learning Lab. Homepage


Recent Submissions rssimage RSS 1.0 RSS 2.0 ATOM 1.0

BROWSE