TimePre: Bridging Accuracy, Efficiency, and Stability in Probabilistic Time-Series Forecasting

TL;DR

TimePre is a unified framework for probabilistic time-series forecasting that combines MLP efficiency with flexible distribution modeling, achieving state-of-the-art accuracy, speed, and stability.

Contribution

It introduces Stabilized Instance Normalization (SIN) to enhance hybrid models, resolving stability issues and improving inference speed and accuracy.

Findings

Achieves SOTA accuracy on six benchmark datasets.

Orders of magnitude faster inference than sampling-based models.

More stable than prior MCL approaches.

Abstract

We propose TimePre, a simple framework that unifies the efficiency of Multilayer Perceptron (MLP)-based models with the distributional flexibility of Multiple Choice Learning (MCL) for Probabilistic Time-Series Forecasting (PTSF). Stabilized Instance Normalization (SIN), the core of TimePre, is a normalization layer that explicitly addresses the trade-off among accuracy, efficiency, and stability. SIN stabilizes the hybrid architecture by correcting channel-wise statistical shifts, thereby resolving the catastrophic hypothesis collapse. Extensive experiments on six benchmark datasets demonstrate that TimePre achieves state-of-the-art (SOTA) accuracy on key probabilistic metrics. Critically, TimePre achieves inference speeds that are orders of magnitude faster than sampling-based models, and is more stable than prior MCL approaches.