AI researchers develop physics-informed transformer for universal building thermal models

Researchers have developed a physics-informed transformer architecture designed to create a universal thermal model for residential buildings. This model embeds domain knowledge and uses Rotary Position Embedding attention to capture temporal dependencies, aiming for generalization across diverse buildings and climates without specific calibration. Evaluated on the CityLearn dataset, the model demonstrated strong prediction accuracy and zero-shot transferability, outperforming existing baselines and foundation models. AI

影响 This research could lead to more efficient building energy management systems by enabling generalized thermal modeling.

排序理由 This is an academic paper presenting a new model architecture and evaluation results on a specific dataset. [lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

• CityLearn
• Rotary Position Embedding

• 论文
• 其他

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →