This published patent application describes a learning framework for inferring user activities from IoT device events. It extracts representative activity patterns and trains a …
This study infers smart-home user activities directly from overlapping IoT network traffic patterns. Wavelet analysis separates activity-related signals from background traffic, …
IoTDuet uses smart-home network traffic to determine whether an IoT device event was triggered locally or remotely. It examines cloud communication by controlling devices, …
This talk discusses how to analyze vulnerabilities in an attack graph using attack traces and ways to harden them.
This paper introduces weighted attack graphs for analyzing IoT vulnerabilities and selecting network protections. It develops a shortest-attack-trace algorithm, proves the …
This talk discusses how IoT networks can be analyzed in weighted attack graphs and how an effective hardening plan can be devised using attack graphs and attack traces.
Device failures can make different smart-home activities produce indistinguishable IoT event patterns. This paper analyzes those limitations in existing inference algorithms and …
This talk discusses how user activities can be inferred from IoT device events even when device failures make activity patterns ambiguous.
This study combines deterministic pattern extraction with unsupervised learning to infer smart-home user activities from IoT device events. Experiments with 2,959 real activities …