Applications
We apply machine learning and statistical inference to network traffic and other environmental signals in networked environments to (1) infer and improve network performance; (2) infer human behavior and activities in a wide range of circumstances (e.g., health, education); and (3) network privacy and security.
Performance
We apply systems-driven network measurement and inference techniques to measure network performance, application quality, and user quality of experience (QoE).
Activity Recognition
We apply machine learning and inference techniques to derive information about human activity from network traffic and other environmental signals.
Security and Privacy
We apply machine learning to understand and mitigate security and privacy threats from our increasingly connected world.
Foundations
We develop techniques for improving network inference, including the exploration of data representation for network modeling problems, efficient models, and operational models for network analytics.
Data Representation
We explore efficient representations for network traffic for both supervised and unsupervised learning problems.
Operational Analytics
We develop systems and models to enable operational network analytics, including cost-sensitive models and model drift.
We publish papers in top-tier networking, security, and machine learning/modeling conferences, also also regularly publish open-source software. An important value of our lab is real-world impact, through the deployment of operational systems. Contact us to learn more about the group, including how to join.