Dr. Sicong Shao
Assistant Professor: University of North Dakota, School of Electrical Engineering and Computer Science
Title: TinyML ADD: Tiny Machine Learning for Adaptive Anomaly Detection
Tiny machine learning (TinyML)-based anomaly detection, bringing anomaly detection, edge computing, and AI together to power-constrained edge devices, has become an emerging research topic. However, it may become suboptimal, unreliable, or even catastrophically fail at worst when concept drift affects the data generation process. This project aims to develop a TinyML-based framework that can effectively deploy robust anomaly detection on edge devices and enable edge AI to perform on-device adaptive learning to deal with non-stationary environments.