# Finite-Horizon Quickest Change Detection Balancing Latency with False Alarm Probability

arXiv: https://arxiv.org/abs/2511.12803 · HTML: https://mlsanity.com/paper/2511.12803/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 21 references that are in the map.
- First posted on arXiv 16 Nov 2025.

## Builds on (most cited first, 5 of 5)

- [Non-stationary Reinforcement Learning without Prior Knowledge: An Optimal Black-box Approach](https://mlsanity.com/paper/2102.05406/) (2021)
- [Mixture Martingales Revisited with Applications to Sequential Tests and Confidence Intervals](https://mlsanity.com/paper/1811.11419/) (2018)
- [A New Look at Dynamic Regret for Non-Stationary Stochastic Bandits](https://mlsanity.com/paper/2201.06532/) (2022)
- [Sequential Change Detection for Learning in Piecewise Stationary Bandit Environments](https://mlsanity.com/paper/2501.10974/) (2025)
- [High Probability Latency Sequential Change Detection over an Unknown Finite Horizon](https://mlsanity.com/paper/2408.05817/) (2024)

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

- [DARLING: Detection Augmented Reinforcement Learning with Non-Stationary Guarantees](https://mlsanity.com/paper/2604.16684/) (2026)

## Abstract (arXiv)

A finite-horizon variant of the quickest change detection (QCD) problem that is of relevance to learning in non-stationary environments is studied. The metric characterizing false alarms is the probability of a false alarm occurring before the horizon ends. The metric that characterizes the delay is latency, which is the smallest value such that the probability that detection delay exceeds this value is upper bounded to a predetermined latency level. The objective is to minimize the latency (at a given latency level), while maintaining a low false alarm probability. Under the pre-specified latency and false alarm levels, a universal lower bound on the latency, which any change detection procedure needs to satisfy, is derived. Change detectors are then developed, which are order-optimal in terms of the horizon. The case where the pre- and post-change distributions are known is considered first, and then the results are generalized to the non-parametric case when they are unknown except that they are sub-Gaussian with different means. Simulations are provided to validate the theoretical results.