The field of battery design and building energy modeling is moving towards increased efficiency and accuracy through the integration of machine learning and physics-guided approaches. Recent developments have focused on reducing the need for prototyping and improving the reliability of predictions, enabling rapid feedback and evaluation of new design candidates. This shift is driven by the need for fast and reliable validation of complex physical systems, such as batteries, and the importance of accurate energy consumption forecasting in building management. Notable papers in this area include:

• Discovery Learning, which introduces a scientific machine-learning paradigm that integrates active learning, physics-guided learning, and zero-shot learning to accelerate battery design evaluation.
• Physics-Guided Memory Network, which integrates predictions from deep learning and physics-based models to address the limitations of each approach in building energy modeling.