Evolve predicts future anomalies in complex time series data effectively
EVOLVE: Uncertainty-Aware Multivariate Time Series Anomaly Prediction with Adaptive Prototype Memories
Machine Learning
Summary
Predicting future problems in systems from current data is hard because the same signs can mean different things depending on the situation. The authors designed EVOLVE, a method that learns from past data to understand how unusual changes develop differently for each part of a system and situation. It guesses multiple possible future paths and uses differences between them to tell if problems might happen soon. Tests on many real-world datasets show EVOLVE predicts issues more accurately than other methods.
What this means in practice
- •For industrial maintenance teams: Identify early signs of equipment failures by analyzing sensor data with context-aware anomaly predictions before issues occur.
- •For finance risk analysts: Improve detection of complex irregular patterns in financial time series by using uncertainty-aware prediction of potential future anomalies.
Authors
Jianxin Jin, Kai Zhao, Shiyan Hu, Yang Shu, Peng Chen, Meng Wang, Zhongwen Rao, Yijie Li, Jing Wang, Bin Yang, Chenjuan Guo
Abstract
Unsupervised multivariate time series anomaly prediction (AP) aims to estimate future anomaly risks from current observations, providing early-warning signals for proactive intervention and maintenance in complex systems. Existing unsupervised AP methods follow a perturbation-based paradigm, where predefined or random perturbations are injected into historical windows to simulate abnormal future evolution. However, these methods share perturbations across different variables and contexts, and learn deterministic history-to-future evolution. In reality, anomaly evolution is variable- and context-dependent: the same perturbation may be amplified into abnormal future dynamics in one variable or context, while being absorbed as normal fluctuation in another. In addition, future evolution is not uniquely determined by a historical window, since similar histories may lead to multiple plausible trajectories, such as gradual degradation or abrupt changes. To address these problems, we propose EVOLVE, a unified framework that models unsupervised AP via variable- and context-conditioned perturbations and uncertainty-aware future reasoning. EVOLVE constructs context-specific anomaly perturbations through an adaptive anomaly prototype memory, which can dynamically update and retrieve relevant perturbations at the variable level. EVOLVE models uncertainty by sampling multiple latent historical states and decoding through complementary smooth and deviation-amplifying evolution hypotheses, and the disagreement among decoded future trajectories is used to estimate future anomaly risk. Extensive experiments on six real-world benchmark datasets demonstrate that EVOLVE consistently outperforms state-of-the-art anomaly prediction baselines.