Privacy-preserving transfer learning framework for building energy forecasting with fully anonymized data

AI-driven forecasting offers a promising solution for optimal building energy control, yet it is constrained by scarce labeled data and strict privacy regulations. While transfer learning (TL) can alleviate data scarcity by leveraging data from other buildings, conventional TL relies on metadata unavailable in fully anonymized datasets. We propose a Privacy-Preserving Transfer Learning (PPTL) framework that overcomes this deadlock by learning similarity directly from anonymized time-series dynamics. Using an unsupervised contrastive encoder, the framework maps each building’s dynamics to high-dimensional representation vectors learned solely from temporal patterns. Cosine distance between representations guides source selection to pretrain a lightweight forecaster, which is then fine-tuned on limited target data. Leave-one-out experiments on 89 real-world buildings validate that learned similarity strongly correlates with transfer performance: models pretrained on highly similar sources achieve median MSE reductions of 27–31%, peaking at 31% with optimal configurations, compared to target-only baselines. The framework improves forecasting in 99.2% of configurations (353 of 356), with only three instances showing marginal degradation (maximum 2.2%). Furthermore, it effectively mitigates negative transfer by identifying optimal source configurations, transforming transfer learning from an unpredictable trial into a manageable engineering problem. Compared to federated learning, PPTL requires only 0.51% of the communication bandwidth while offloading all computation to the server, enabling deployment on legacy building systems. The framework offers a practical, scalable solution for privacy-preserving collaboration on anonymized building energy data, supporting broader AI adoption in the building sector and potentially in other privacy-constrained time-series domains.