# Conditional Score Learning for Quickest Change Detection in Markov Transition Kernels

arXiv: https://arxiv.org/abs/2511.03953 · HTML: https://mlsanity.com/paper/2511.03953/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 29 references that are in the map.
- First posted on arXiv 6 Nov 2025.

## Builds on (most cited first, 7 of 7)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Score-Based Generative Modeling through Stochastic Differential Equations](https://mlsanity.com/paper/2011.13456/) (2020)
- [Generative Modeling by Estimating Gradients of the Data Distribution](https://mlsanity.com/paper/1907.05600/) (2019)
- [Quickest Change Detection for Unnormalized Statistical Models](https://mlsanity.com/paper/2302.00250/) (2023)
- [Change Detection of Markov Kernels with Unknown Pre and Post Change Kernel](https://mlsanity.com/paper/2201.11722/) (2022)
- [Robust Score-Based Quickest Change Detection](https://mlsanity.com/paper/2407.11094/) (2024)
- [Large Deviation Analysis of Score-based Hypothesis Testing](https://mlsanity.com/paper/2401.15519/) (2024)

## Built on by (most cited first, 1 of 1)

- [Score-Based Change-Point Detection and Region Localization for Spatio-Temporal Point Processes](https://mlsanity.com/paper/2602.04798/) (2026)

## Abstract (arXiv)

We address the problem of quickest change detection in Markov processes with unknown transition kernels. The key idea is to learn the conditional score nablay log p(y|x) directly from sample pairs ( x,y), where both x and y are high-dimensional data generated by the same transition kernel. In this way, we avoid explicit likelihood evaluation and provide a practical way to learn the transition dynamics. Based on this estimation, we develop a score-based CUSUM procedure that uses conditional Hyvarinen score differences to detect changes in the kernel. To ensure bounded increments, we propose a truncated version of the statistic. With Hoeffding's inequality for uniformly ergodic Markov processes, we prove exponential lower bounds on the mean time to false alarm. We also prove asymptotic upper bounds on detection delay. These results give both theoretical guarantees and practical feasibility for score-based detection in high-dimensional Markov models.