Kailai Sun, Zijie Gao, Mingyi He, Can Rong, Alok Prakash, Baoshen Guo, Shenhao Wang, Jinhua Zhao (Massachusetts Institute of Technology) · Urban Energy GenAI

DySENSE

DySENSE is a multimodal generative AI framework that synthesises static urban satellite imagery and spatially aligned, dynamic monthly building energy consumption maps, separating long-term urban morphology from short-term energy use. Across New York, Washington, Norfolk and Busan, the authors report annotated synthetic energy data built from less than 25% labelled data, an 8% mIoU gain in downstream prediction, and 9%-18% lower CVRMSE (ASHRAE) error than existing methods. The data (MUSE2 on Hugging Face) and code (GitHub) are linked from the paper.

Open the source ↗

Similar Research & Safety releases

Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.