One-Embedding-Fits-All: Efficient Zero-Shot Time Series Forecasting by a Model Zoo

Hao-Nan Shi, Ting-Ji Huang, Lu Han, De-Chuan Zhan, Han-Jia Ye

Keywords: Time series forecasting, Model Zoo, Zero-shot forecasting, Foundation models, Model Selection, Model-task co-embedding

TL;DR: For unseen time series zero-shot forecasting task, model-task co-embedding enables O(1)-complexity selection of optimal foundation models from a growing model zoo.

Abstract: The proliferation of Time Series Foundation Models (TSFMs) has significantly advanced zero-shot forecasting, enabling predictions for unseen time series without task-specific fine-tuning. Extensive research has confirmed that no single TSFM excels universally, as different models exhibit preferences for distinct temporal patterns. This diversity suggests an opportunity: how to take advantage of the complementary abilities of TSFMs. To this end, we propose ZooCast, which characterizes each model's distinct forecasting strengths. ZooCast can intelligently assemble current TSFMs into a model zoo that dynamically selects optimal models for different forecasting tasks. Our key innovation lies in the One-Embedding-Fits-All paradigm that constructs a unified representation space where each model in the zoo is represented by a single embedding, enabling efficient similarity matching for all tasks. Experiments demonstrate ZooCast's strong performance on the GIFT-Eval zero-shot forecasting benchmark while maintaining the efficiency of a single TSFM. In real-world scenarios with sequential model releases, the framework seamlessly adds new models for progressive accuracy gains with negligible overhead.

Primary Area: learning on time series and dynamical systems

Submission Number: 7264