Differentiable Community Detection with GNNs and Stochastic Block Models

Using SBM likelihood as loss functions for unsupervised GNN training

Using Hypervectors for Efficient Anomaly Detection in Graph Streams

Real-time network anomaly detection with Hyperdimensional Computing

Using Machine Learning to Predict Critical Security Vulnerabilities from Twitter

Our NAACL 2019 paper on analyzing cybersecurity threats on social media was featured in Wired

Boxing's Grandmasters Part II

Analysis of Historical boxing data

Deep Learning for Time Series, pt. 1