Time-o1: Time-Series Forecasting Needs Transformed Label Alignment

2025/05/23 by Hao Wang, Wang, Hao, Zhichao Chen +12 · 8 citations
Computer Science · #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #FOS: Electrical engineering #Machine Learning (cs.LG) #Systems and Control (eess.SY) #Time Series Analysis and Forecasting #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2505.17847

openalex publication_date 2025/05/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Training time-series forecast models presents unique challenges in designing effective learning objectives. Existing methods predominantly utilize the temporal mean squared error, which faces two critical challenges: (1) label autocorrelation, which leads to bias from the label sequence likelihood; (2) excessive amount of tasks, which increases with the forecast horizon and complicates optimization. To address these challenges, we propose Time-o1, a transformation-augmented learning objective tailored for time-series forecasting. The central idea is to transform the label sequence into decorrelated components with discriminated significance. Models are then trained to align the most significant components, thereby effectively mitigating label autocorrelation and reducing task amount. Extensive experiments demonstrate that Time-o1 achieves state-of-the-art performance and is compatible with various forecast models. Code is available at https://github.com/Master-PLC/Time-o1.

Cited by

• Weighted Contrastive Learning for Anomaly-Aware Time-Series Forecasting
• DistDF: Time-Series Forecasting Needs Joint-Distribution Wasserstein Alignment
• Quadratic Direct Forecast for Training Multi-Step Time-Series Forecast Models
• DBLoss: Decomposition-based Loss Function for Time Series Forecasting
• Enhancing Time Series Forecasting through Selective Representation Spaces: A Patch Perspective
• ASTGI: Adaptive Spatio-Temporal Graph Interactions for Irregular Multivariate Time Series Forecasting
• Aurora: Towards Universal Generative Multimodal Time Series Forecasting
• Bridging Past and Future: Distribution-Aware Alignment for Time Series Forecasting

Related

• Bridging Past and Future: Distribution-Aware Alignment for Time Series Forecasting
• LangTime: A Language-Guided Unified Model for Time Series Forecasting with Proximal Policy Optimization
• DDTime: Dataset Distillation with Spectral Alignment and Information Bottleneck for Time-Series Forecasting
• FrAug: Frequency Domain Augmentation for Time Series Forecasting
• DistDF: Time-Series Forecasting Needs Joint-Distribution Wasserstein Alignment