Bryan Hooi

Assistant Professor
National University of Singapore

E-mail: bhooi@comp.nus.edu.sg
Google Scholar | Twitter


I'm an Assistant Professor in the School of Computing and the Institute of Data Science in National University of Singapore. My research aims to make machine learning systems more reliable and applicable to a wider variety of real-world contexts, particularly:

I am actively recruiting postdoctoral fellows as well as NUS students of all levels interested in the above topics. Please email me if you are interested.

I received my Ph.D. in Machine Learning from Carnegie Mellon University, where I was advised by Christos Faloutsos. I received my M.S. in Computer Science and my B.S. in Mathematics from Stanford University.



Branch and Border: Partition-Based Change Detection in Multivariate Time Series.
BNB (Branch and Border) is a nonparametric multivariate change detection method. BNB approaches change detection by separating points before and after the change using an ensemble of random partitions.

BeatLex: Summarizing and Forecasting Time Series with Patterns.
BeatLex is an algorithm that succintly summarizes and forecasts time series data. It is designed for data containing patterns that occur repeatedly, especially if these patterns are complex and nonlinear, change over time, and may distortions in their shape or length.

FRAUDAR: Bounding Graph Fraud in the Face of Camouflage.
FRAUDAR is an algorithm for detecting graph fraud based on dense subgraph detection, which is aimed at being robust to camouflage (i.e. attackers which add false edges in order to mask their presence)

BIRDNEST: Bayesian Inference for Ratings-Fraud Detection
BIRDNEST is an algorithm for detecting fraudulent users in timestamped ratings data (e.g. users ratings products with 1 to 5 stars), based on detecting users that differ excessively from the norm, in terms of what ratings they give and the time distribution of their ratings.

Matrices, Compression, Learning Curves: formulation, and the GROUPNTEACH algorithms.
GNT takes a collection of facts, arranged in a binary matrix, and reorders the rows and columns for the purpose of teaching and visualization.