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

2025-09-04

AI Summary

Single time-series foundation models (TSFMs) struggle to achieve universal optimality across diverse forecasting tasks. Method: This paper proposes ZooCast—a framework that constructs a “zoo” of TSFMs and introduces the novel One-Embedding-Fits-All paradigm, mapping heterogeneous models into a unified embedding space. This enables zero-shot dynamic model selection and lightweight ensemble inference via model embedding learning and similarity-based matching. Contribution/Results: ZooCast supports efficient retrieval, seamless incremental model integration, and cross-task generalization. Evaluated on the GIFT-Eval benchmark, it significantly outperforms individual TSFMs and state-of-the-art ensemble methods—achieving consistent accuracy gains while preserving single-model inference latency and enabling scalable deployment.

Technology Category

• Machine Learning: Ensemble Methods
• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

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.

Problem

• Leveraging diverse Time Series Foundation Models' complementary strengths
• Dynamically selecting optimal models for zero-shot forecasting tasks
• Creating unified embedding space for efficient model-task matching

Innovation

• Model zoo with dynamic model selection
• One-Embedding-Fits-All unified representation space
• Efficient similarity matching across forecasting tasks