Research

I am interested in a wide range of topics in statistics. I used to work on these topics:

• Detection and localization of anomalies;

• Clustering algorithms analysis and design;

• Robust and nonparametric methods, such as rankings;

During my work in industry, I am focusing on:

• Computational ads;

• Anomaly detection with robustness and accuracy;

• Scientific AB testing and algorithm evaluation;

I review for venues such as AISTATS, ICML and NeurIPS.

Publications:

Distribution-free Detection of a Submatrix

E. Arias-Castro, Y. Liu, Journal of Multivariate Analysis, 2017.

Distribution-Free, Size Adaptive Submatrix Detection with Acceleration

Y.Liu and J.Guo, ALEA, Latin American Journal of Probability and Mathematical Statistics, 2020.

Finding Multidimensional Patterns in Multidimensional Time Series

E.Laftchiev and Y.Liu, SIGKDD Workshop on Mining and Learning From Time Series, 2018.

A Multiscale Scan Statistic for Adaptive Submatrix Localization

Y.Liu and E. Arias-Castro, ACM SIGKDD Research Track Paper, 2019.