AsyTO: Asymmetric Temporal Operator for Parameter-Efficient Multivariate Time Series Forecasting

2026-08-17

AI Summary

This study addresses the inherent trade-off between parameter efficiency and modeling flexibility in multivariate time series forecasting by proposing AsyTO, an asymmetric temporal operator. By decoupling historical reading from future generation patterns and incorporating low-rank periodic prototypes with separable factorization mechanisms, AsyTO shifts compression from observed sequences to predictive operators, enabling parameter complexity to scale linearly with the number of variables. Extensive evaluations across 44 configurations on 11 benchmarks demonstrate that AsyTO achieves state-of-the-art lightweight performance in 30 settings, successfully positioning itself on the accuracy-computation Pareto frontier. Consequently, this work establishes a novel paradigm for efficient multivariate forecasting that effectively reconciles model compactness with predictive capability.

Abstract

Multivariate time-series forecasting faces a structural dilemma: sharing one temporal predictor across variables is parameter-efficient but forces heterogeneous variables through an identical history-to-future map, whereas learning an independent predictor per variable restores flexibility at a cost that grows with the product of variable count, context length, and horizon. We argue that this dilemma dissolves once the object being compressed is the forecasting operator rather than the observed series. Auditing per-variable linear history-to-future maps across standard benchmarks, we find that a phase-locked seasonal component paired with a compact residual operator outperforms a dense phase-blind reference in most audited settings. The residual transport is also directional: lag-invariant alternatives consistently underperform asymmetric history-to-future maps. Guided by this structure, we propose AsyTO, an Asymmetric Temporal Operator that factorizes the tensor of per-variable operators into shared but distinct history-reading and future-writing temporal modes with per-variable mode-wise gains, complemented by a low-rank periodic prototype and a cycle-separable factorization of the temporal modes. Each forecast reads only its own variable's history, so parameters and compute grow linearly in the number of variables. Across eleven benchmarks and multiple forecast horizons, AsyTO attains the best lightweight error in 30 of 44 dataset-horizon settings, locating at the accuracy-compute Pareto frontier.

Problem

• Multivariate Time Series Forecasting
• Parameter Efficiency
• Heterogeneous Variables
• Temporal Operator

Innovation

• Asymmetric Temporal Operator
• Parameter-Efficient Forecasting
• Tensor Factorization
• Cycle-Separable Factorization
• Multivariate Time Series