Black-Box Time-Series Domain Adaptation via Cross-Prompt Foundation Models

2025-10-01

AI Summary

This paper addresses black-box time-series domain adaptation (BBTSDA)—a novel problem where existing black-box domain adaptation methods, designed primarily for vision tasks, fail to capture the spatiotemporal dynamics inherent in time-series data, and the potential of foundation models in this setting remains unexplored. To tackle this, we propose Cross-Prompt Foundation Models (CPFM), the first framework to integrate foundation models into BBTSDA. CPFM employs a dual-branch architecture and a two-level reconstruction learning mechanism—operating jointly at the prompt level and input level—to enable privacy-preserving cross-domain transfer using only source-model API queries. Experiments on three cross-domain time-series benchmarks demonstrate that CPFM significantly outperforms state-of-the-art black-box methods, validating its effectiveness, generalizability, and capacity to model spatiotemporal characteristics.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Diffusion Models for Vision
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

The black-box domain adaptation (BBDA) topic is developed to address the privacy and security issues where only an application programming interface (API) of the source model is available for domain adaptations. Although the BBDA topic has attracted growing research attentions, existing works mostly target the vision applications and are not directly applicable to the time-series applications possessing unique spatio-temporal characteristics. In addition, none of existing approaches have explored the strength of foundation model for black box time-series domain adaptation (BBTSDA). This paper proposes a concept of Cross-Prompt Foundation Model (CPFM) for the BBTSDA problems. CPFM is constructed under a dual branch network structure where each branch is equipped with a unique prompt to capture different characteristics of data distributions. In the domain adaptation phase, the reconstruction learning phase in the prompt and input levels is developed. All of which are built upon a time-series foundation model to overcome the spatio-temporal dynamic. Our rigorous experiments substantiate the advantage of CPFM achieving improved results with noticeable margins from its competitors in three time-series datasets of different application domains.

Problem

• Addressing black-box domain adaptation for time-series data
• Leveraging foundation models for spatio-temporal characteristics
• Developing cross-prompt framework for improved domain adaptation

Innovation

• Cross-Prompt Foundation Model for time-series adaptation
• Dual branch network with unique prompt design
• Reconstruction learning at prompt and input levels