Score-based Quickest Change Detection for Unnormalized Models

Suya Wu, Enmao Diao, Taposh Banerjee, Jie Ding, Vahid Tarokh
Proceedings of The 26th International Conference on Artificial Intelligence and Statistics, PMLR 206:10546-10565, 2023.

Abstract

Classical change detection algorithms typically require modeling pre-change and post-change distributions. The calculations may not be feasible for various machine learning models because of the complexity of computing the partition functions and normalized distributions. Additionally, these methods may suffer from a lack of robustness to model mismatch and noise. In this paper, we develop a new variant of the classical Cumulative Sum (CUSUM) change detection, namely Score-based CUSUM (SCUSUM), based on Fisher divergence and the Hyvärinen score. Our method allows the applications of the quickest change detection for unnormalized distributions. We provide a theoretical analysis of the detection delay given the constraints on false alarms. We prove the asymptotic optimality of the proposed method in some particular cases. We also provide numerical experiments to demonstrate our method’s computation, performance, and robustness advantages.

Cite this Paper


BibTeX
@InProceedings{pmlr-v206-wu23b, title = {Score-based Quickest Change Detection for Unnormalized Models}, author = {Wu, Suya and Diao, Enmao and Banerjee, Taposh and Ding, Jie and Tarokh, Vahid}, booktitle = {Proceedings of The 26th International Conference on Artificial Intelligence and Statistics}, pages = {10546--10565}, year = {2023}, editor = {Ruiz, Francisco and Dy, Jennifer and van de Meent, Jan-Willem}, volume = {206}, series = {Proceedings of Machine Learning Research}, month = {25--27 Apr}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v206/wu23b/wu23b.pdf}, url = {https://proceedings.mlr.press/v206/wu23b.html}, abstract = {Classical change detection algorithms typically require modeling pre-change and post-change distributions. The calculations may not be feasible for various machine learning models because of the complexity of computing the partition functions and normalized distributions. Additionally, these methods may suffer from a lack of robustness to model mismatch and noise. In this paper, we develop a new variant of the classical Cumulative Sum (CUSUM) change detection, namely Score-based CUSUM (SCUSUM), based on Fisher divergence and the Hyvärinen score. Our method allows the applications of the quickest change detection for unnormalized distributions. We provide a theoretical analysis of the detection delay given the constraints on false alarms. We prove the asymptotic optimality of the proposed method in some particular cases. We also provide numerical experiments to demonstrate our method’s computation, performance, and robustness advantages.} }
Endnote
%0 Conference Paper %T Score-based Quickest Change Detection for Unnormalized Models %A Suya Wu %A Enmao Diao %A Taposh Banerjee %A Jie Ding %A Vahid Tarokh %B Proceedings of The 26th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2023 %E Francisco Ruiz %E Jennifer Dy %E Jan-Willem van de Meent %F pmlr-v206-wu23b %I PMLR %P 10546--10565 %U https://proceedings.mlr.press/v206/wu23b.html %V 206 %X Classical change detection algorithms typically require modeling pre-change and post-change distributions. The calculations may not be feasible for various machine learning models because of the complexity of computing the partition functions and normalized distributions. Additionally, these methods may suffer from a lack of robustness to model mismatch and noise. In this paper, we develop a new variant of the classical Cumulative Sum (CUSUM) change detection, namely Score-based CUSUM (SCUSUM), based on Fisher divergence and the Hyvärinen score. Our method allows the applications of the quickest change detection for unnormalized distributions. We provide a theoretical analysis of the detection delay given the constraints on false alarms. We prove the asymptotic optimality of the proposed method in some particular cases. We also provide numerical experiments to demonstrate our method’s computation, performance, and robustness advantages.
APA
Wu, S., Diao, E., Banerjee, T., Ding, J. & Tarokh, V.. (2023). Score-based Quickest Change Detection for Unnormalized Models. Proceedings of The 26th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 206:10546-10565 Available from https://proceedings.mlr.press/v206/wu23b.html.

Related Material