conftrace_

2023 AISTATS AISTATS 2023

A Contrastive Approach to Online Change Point Detection

Abstract

We suggest a novel procedure for online change point detection. Our approach expands an idea of maximizing a discrepancy measure between points from pre-change and post-change distributions. This leads to a flexible procedure suitable for both parametric and nonparametric scenarios. We prove non-asymptotic bounds on the average running length of the procedure and its expected detection delay. The efficiency of the algorithm is illustrated with numerical experiments on synthetic and real-world data sets.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Data Science & Analytics

Authors

Nikita Puchkin, Valeriia Shcherbakova

Topics

Machine Learning > Core Methods > Classification Machine Learning > Learning Types > Contrastive Learning Machine Learning > Learning Types > Unsupervised Learning Machine Learning > Optimization & Theory > Statistical Learning Data Science & Analytics > Methods > Time Series Deep Learning > Learning Types > Contrastive Learning

Keywords

change point detection contrastive learning distribution shift online algorithm statistical bound non-parametric method online detection nonparametric detection

More from AISTATS 2023

All AISTATS 2023 papers →

Faithful Heteroscedastic Regression with Neural Networks

Boosted Off-Policy Learning

Symmetric (Optimistic) Natural Policy Gradient for Multi-Agent Learning with Parameter Convergence

Active Membership Inference Attack under Local Differential Privacy in Federated Learning

Differentially Private Matrix Completion through Low-rank Matrix Factorization

Private Non-Convex Federated Learning Without a Trusted Server