Zeus: Towards Tuning-Free Foundation Model for Time Series Analysis

2026-07-02

AI Summary

This work proposes a unified, zero-shot time series foundation model that overcomes the reliance on task-specific fine-tuning commonly observed in existing approaches. By integrating a multi-scale Transformer architecture—combining point-level tokenization with a U-shaped hierarchical structure—the model effectively balances fine-grained temporal modeling and computational efficiency for long sequences. Furthermore, it introduces a Multi-Objective Temporal Masking (MOTM) mechanism to uniformly handle heterogeneous tasks such as forecasting, imputation, and global abstraction within a single framework. Evaluated across five representative time series tasks, the model demonstrates strong performance without any fine-tuning, underscoring its effectiveness and potential as a general-purpose foundation model for diverse time series applications.

Abstract

We present Zeus, a unified tuning-free Time Series Foundation Model (TSFM) that delivers superior performance across diverse analysis tasks without any task-specific fine-tuning. Unlike prior studies that primarily focus on zero-shot forecasting but require task-specific tuning for other tasks, Zeus bridges this gap by addressing two fundamental challenges in multi-task generalization. First, to reconcile point-level granularity with long-sequence scalability, Zeus incorporates a multi-scale Transformer featuring point-wise tokenization and a U-shaped hierarchy, effectively balancing fine-grained fidelity with computational efficiency. Second, to accommodate varying inductive biases across different tasks, Zeus introduces Multi-Objective Temporal Masking (MOTM), a unified strategy that supports heterogeneous tasks (e.g., extrapolation, interpolation, and global abstraction) within a single framework. Extensive experiments across five representative tasks demonstrate that Zeus consistently achieves competitive results in tuning-free settings, underscoring its potential as a general-purpose TSFM.

Problem

• Time Series Foundation Model
• multi-task generalization
• tuning-free
• zero-shot forecasting
• inductive biases

Innovation

• Tuning-Free
• Time Series Foundation Model
• Multi-Scale Transformer
• Multi-Objective Temporal Masking
• Multi-Task Generalization