OneCast: Structured Decomposition and Modular Generation for Cross-Domain Time Series Forecasting

2025-10-27

AI Summary

To address trend drift and inconsistent periodicity arising from heterogeneous data in cross-domain time-series forecasting, this paper proposes OneCast—a novel framework grounded in structural disentanglement. It introduces a semantic-aware tokenizer coupled with a masked discrete diffusion model to explicitly capture non-stationary trend evolution, and employs a basis-function-driven lightweight projection module for interpretable and robust periodic reconstruction. The dual-branch architecture jointly optimizes interpretability and generalization. Evaluated across eight diverse domains—including energy, transportation, and healthcare—OneCast consistently outperforms state-of-the-art methods, achieving significant gains in both prediction accuracy and cross-domain stability. Its modular design enables effective transfer across disparate data distributions without domain-specific adaptation, demonstrating superior robustness to distributional shifts and heterogeneity.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

Cross-domain time series forecasting is a valuable task in various web applications. Despite its rapid advancement, achieving effective generalization across heterogeneous time series data remains a significant challenge. Existing methods have made progress by extending single-domain models, yet often fall short when facing domain-specific trend shifts and inconsistent periodic patterns. We argue that a key limitation lies in treating temporal series as undifferentiated sequence, without explicitly decoupling their inherent structural components. To address this, we propose OneCast, a structured and modular forecasting framework that decomposes time series into seasonal and trend components, each modeled through tailored generative pathways. Specifically, the seasonal component is captured by a lightweight projection module that reconstructs periodic patterns via interpretable basis functions. In parallel, the trend component is encoded into discrete tokens at segment level via a semantic-aware tokenizer, and subsequently inferred through a masked discrete diffusion mechanism. The outputs from both branches are combined to produce a final forecast that captures seasonal patterns while tracking domain-specific trends. Extensive experiments across eight domains demonstrate that OneCast mostly outperforms state-of-the-art baselines.

Problem

• Addressing generalization challenges in cross-domain time series forecasting
• Overcoming domain-specific trend shifts and inconsistent periodic patterns
• Decomposing time series into seasonal and trend components separately

Innovation

• Decomposes time series into seasonal and trend components
• Uses interpretable basis functions for seasonal patterns
• Employs masked discrete diffusion for trend modeling