Bridging Past and Future: Distribution-Aware Alignment for Time Series Forecasting
This addresses the problem of distributional misalignment in time series forecasting for practitioners, offering a novel method that improves performance on existing benchmarks.
The paper tackles the limited adoption of representation-learning methods in time series forecasting by introducing TimeAlign, a plug-and-play framework that aligns past and future representations to bridge distributional gaps. Experiments across eight benchmarks show superior performance, with gains attributed to correcting frequency mismatches between inputs and outputs.
Although contrastive and other representation-learning methods have long been explored in vision and NLP, their adoption in modern time series forecasters remains limited. We believe they hold strong promise for this domain. To unlock this potential, we explicitly align past and future representations, thereby bridging the distributional gap between input histories and future targets. To this end, we introduce TimeAlign, a lightweight, plug-and-play framework that establishes a new representation paradigm, distinct from contrastive learning, by aligning auxiliary features via a simple reconstruction task and feeding them back into any base forecaster. Extensive experiments across eight benchmarks verify its superior performance. Further studies indicate that the gains arise primarily from correcting frequency mismatches between historical inputs and future outputs. Additionally, we provide two theoretical justifications for how reconstruction improves forecasting generalization and how alignment increases the mutual information between learned representations and predicted targets. The code is available at https://github.com/TROUBADOUR000/TimeAlign.