arXiv:2511.06634v3 Announce Type: replace 
Abstract: Cross-domain HVAC energy prediction is essential for scalable building energy management, particularly because collecting extensive labeled data for every new building is both costly and impractical. Yet, this task remains highly challenging due to the scarcity and heterogeneity of data across different buildings, climate zones, and seasonal patterns. In particular, buildings situated in distinct climatic regions introduce variability that often leads existing methods to overfit to spurious correlations, rely heavily on expert intervention, or compromise on data diversity. To address these limitations, we propose CaberNet, a causal and interpretable deep sequence model that learns invariant (Markov blanket) representations for robust cross-domain prediction. In a purely data-driven fashion and without requiring any prior knowledge, CaberNet integrates i) a global feature gate trained with a self-supervised Bernoulli regularization to distinguish superior causal features from inferior ones, and ii) a domain-wise training scheme that balances domain contributions, minimizes cross-domain loss variance, and promotes latent factor independence. We evaluate CaberNet on real-world datasets collected from three buildings located in three climatically diverse cities, and it consistently outperforms all baselines, achieving a 22.9% reduction in normalized mean squared error (NMSE) compared to the best benchmark. Our code is available at https://github.com/SusCom-Lab/CaberNet-CRL.

تم تقديم نموذج جديد يسمى CaberNet للتنبؤ بالطاقة في أنظمة HVAC عبر المجالات، حيث يعالج التحديات المتعلقة بنقص البيانات وتنوعها عبر المباني والمناخات المختلفة. يستخدم هذا النموذج نهجًا عميقًا تسلسليًا سببيًا وقابلًا للتفسير لتعلم تمثيلات ثابتة، مما يعزز دقة التنبؤات دون الحاجة إلى بيانات موسعة أو معرفة مسبقة.

Se ha presentado un nuevo modelo llamado CaberNet para la predicción de energía HVAC entre dominios, abordando los desafíos relacionados con la escasez y variabilidad de datos en diferentes edificios y climas. Este modelo utiliza un enfoque de secuencia profunda causal e interpretable para aprender representaciones invariantes, mejorando la precisión de las predicciones sin requerir datos etiquetados extensos o conocimiento previo.

Un nouveau modèle nommé CaberNet a été introduit pour la prédiction de l'énergie HVAC inter-domaines, répondant aux défis liés à la rareté et à la variabilité des données entre différents bâtiments et climats. Ce modèle utilise une approche séquentielle profonde causale et interprétable pour apprendre des représentations invariantes, améliorant ainsi la précision des prédictions sans nécessiter de données étiquetées étendues ou de connaissances préalables.

A new model named CaberNet has been introduced for cross-domain HVAC energy prediction, addressing challenges related to data scarcity and variability across different buildings and climates. This model utilizes a causal and interpretable deep sequence approach to learn invariant representations, enhancing prediction accuracy without requiring extensive labeled data or prior knowledge.

CaberNet: Causal Representation Learning for Cross-Domain HVAC Energy Prediction

One More Thing in AI – Your Shortcut to AI Mastery

CaberNet: Causal Representation Learning for Cross-Domain HVAC Energy Prediction

Was this article worth reading? Share it

One More Thing in AI

LucidQuery AI

Airparser

Dyad

AIPortalX

Meteoria

Ready to build your own newsroom?