Change-Point Detection
Overview
Change-point detection studies when and where a time series, data stream, or evolving system shifts between regimes, helping people respond to failures, drift, or changing conditions. Offline methods segment data using dynamic programming and penalized costs, while online methods such as CUSUM, EWMA, and martingale tests trade detection speed against false alarms; Bayesian, kernel, graph-based, and neural approaches handle richer data. Work spans settings from multivariate monitoring and dynamic networks to deployed machine-learning models, with growing attention to confidence sets, attribution, and early-warning evaluation. A central challenge is maintaining reliable guarantees under dependence, adaptation, and unknown shifts without making detection too slow or too prone to alarms.
Latest papers
89 papers9 papers in the last four weeks, down 10% on the four weeks before. 0.1% of all new papers.
Time Series Change Point Detection with Gaussian Processes
TSB-SEG: A Systematic Time-Series Segmentation Benchmark
Shaped by What's Missing: Topological Invariance Simplification Discovers Behavioral Transitions
The First-Order Value of Free Covariates in Budgeted Risk Monitoring
Lightweight Change Detection for Adaptive Collective Perception in Dynamic Environments
Learning from Automatic 3D Gaussian Splatting Edits for Online Bayesian Scene Change Detection
Time-partitioning Gaussian Process Mixture Model
Many Steps into One: Reasoning Contraction and the Limits of Safe Memory Reuse for LLM Agents
Learning to Detect Shifts in Unstructured Data Streams via Monitoring-by-Betting
Monitoring for Shifts at Unknown Times: Parameter-Free Conditional Conformal Martingales
Quickest Change Detection with Diffusion-Integrated Scores
Logarithmic Regret via Passive Change Detection in Piecewise-Stationary Self-Tuning Regulation
CUSP: CUSUM-Governed Survival Hazard Alarms at the Perception Onset for Off-Road Navigation
The Blind Spot Paradox: When Adaptive Classifiers Defeat Drift Detectors
When Is Enough Enough in Self-Evolving LLM Systems?
Beyond Sub-Gaussian Detector Scores: Robust Weighted Profile-Loss Change Point Detection for Human-LLM Text Segmentation
LA-CPD: Local-Evidence-Aware Change-Point Detection for Human-LLM Authorship Segmentation
Period Segmentation in Transition Network Analysis: A Topological Data Analysis Approach
High-Dimensional Online Change Point Detection with Adaptive Thresholding and Interpretability
The Price of Self-Calibration: Exact Evidence Budgets and Manufactured Blind Sets in Adaptive Monitoring
TTDF: A Two-Stage Framework for Reliable Surgical Phase Transition Detection
Distribution-free inference on the number of changepoints
Online Change-point Detection for Cooperative Multi-Agent Reinforcement Learning
Beyond Stationarity in Time Series: Discovering Causal Structures and Latent Regimes via Markov Blankets
When the Martingale Never Stops Firing: Anytime-Valid Gating on Real Forecast Streams
Deciding When to Decide: Testing Operational Suboptimality Under Distributional Shift
RADAR (Regret-based Assessment of Decision Adequacy and Risk), a decision-focused framework that uses inverse optimization to infer latent preferences and tests the deployed decision's optimality gap under the current distribution. By targeting regret, RADAR ignores decision-irrelevant shifts while detecting changes that warrant re-optimization. We develop two-sample and sequential changepoint procedures and establish asymptotic guarantees for Type-I error and power. Across synthetic optimization problems, a semi-synthetic capacity allocation task, and police-zone planning, RADAR more reliably distinguishes harmful from harmless shifts than decision-agnostic alternatives.