Quantum quickest change detection
Quickest change detection (QCD) is a fundamental statistical framework for identifying abrupt anomalies in streaming data. Translating this framework to the quantum domain introduces new challenges. In this paper, we propose a unified optimization framework for QCD of quantum states and quantum channels under varying degrees of prior knowledge. Specifically, for QCD of quantum states, when the post-change state is known, we show that the optimal positive operator-valued measure (POVM) that maximizes the measured Kullback–Leibler (KL) divergence between pre- and post-change states is von Neumann, and develop a projected gradient ascent algorithm that computes a locally optimal von Neumann POVM. When the post-change state is unknown, we show that the maximum-sensitivity POVM is given by the eigen-projectors corresponding to the positive eigenvalues of the pre-change state, and propose a window-based CUSUM-like test. We show that the rank-deficient pre-change state substantially reduces the detection delay via null-space projection. Based on this property, for QCD of quantum channels, we jointly optimize the probe state and the POVM, with the primary objective of inducing rank-deficiency of the channel output state, and the secondary objective of maximizing the KL divergence (for known post-change channel) or the sensitivity (for unknown post-change channel).
Authors
- Jeremy Johnston (ORCID: https://orcid.org/0000-0002-1328-6621)
- Luca Venturino (ORCID: https://orcid.org/0000-0002-7742-4835)
- Zhongyuan Fu
- Xiaodong Wang
Institutions
- Università degli studi di Cassino e del Lazio Meridionale (IT)
- Columbia University (US)
Publication Details
- Journal
- npj Quantum Information
- Published
- 2026-11-17
- DOI
- https://doi.org/10.1038/s41534-026-01418-5
- Primary Topic
- Quantum Information and Cryptography
- Type
- article
- Field-Weighted Citation Impact
- 0.00