A Unified Shape-Aware Foundation Model for Time Series Classification

2026-01-10

arXiv.org

AI Summary

This work proposes UniShape, the first unified shape-aware foundation model for time series classification, addressing the limited modeling of interpretable shapelet features in existing time series foundation models that predominantly focus on forecasting. UniShape introduces a shape-aware adapter to adaptively aggregate multi-scale discriminative subsequences and incorporates a prototype-driven pretraining mechanism to jointly learn transferable representations at both instance and shape levels. Pretrained on a large-scale multi-domain dataset comprising 1.89 million samples, UniShape achieves state-of-the-art performance across 128 UCR datasets and 30 additional benchmarks, significantly enhancing interpretability and cross-domain generalization capabilities.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Computer Vision: Other Foundations of Computer Vision
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Foundation models pre-trained on large-scale source datasets are reshaping the traditional training paradigm for time series classification. However, existing time series foundation models primarily focus on forecasting tasks and often overlook classification-specific challenges, such as modeling interpretable shapelets that capture class-discriminative temporal features. To bridge this gap, we propose UniShape, a unified shape-aware foundation model designed for time series classification. UniShape incorporates a shape-aware adapter that adaptively aggregates multiscale discriminative subsequences (shapes) into class tokens, effectively selecting the most relevant subsequence scales to enhance model interpretability. Meanwhile, a prototype-based pretraining module is introduced to jointly learn instance- and shape-level representations, enabling the capture of transferable shape patterns. Pre-trained on a large-scale multi-domain time series dataset comprising 1.89 million samples, UniShape exhibits superior generalization across diverse target domains. Experiments on 128 UCR datasets and 30 additional time series datasets demonstrate that UniShape achieves state-of-the-art classification performance, with interpretability and ablation analyses further validating its effectiveness.

Problem

• time series classification
• shapelets
• foundation model
• interpretable features
• class-discriminative patterns

Innovation

• shape-aware
• time series classification
• foundation model
• interpretable shapelets
• prototype-based pretraining