BuildFL (Building Federated) is a HVAC Analytic Platform through Federated Learning. With BuildFL, we can easily test a large number of different models on different dataset.

BuildFL is published as a note paper in The Eleventh ACM International Conference on Future Energy Systems (ACM e-Energy '20): Towards Federated Learning for HVAC Analytics: A Measurement Study.

BuildFL provide an abstraction API of model training, intermediate model update, global model distribute in federated learning and support machine learning models used in HVAC analytics.

The document of BuildFL is placed in document\.

BuildFL is written in Python, for now you can use python run.py to stimulate a federated learning training process. Another version of BuildFL which can execute on actual parameter server and participants' device will be published as a python library.

You are welcome to cite our research paper:

@inproceedings{10.1145/3396851.3397717,
author = {Yunzhe Guo, Dan Wang, Arun Vishwanath, Cheng Xu and Qi Li},
title = {Towards Federated Learning for HVAC Analytics: A Measurement Study},
year = {2020},
isbn = {9781450380096},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3396851.3397717},
doi = {10.1145/3396851.3397717},
booktitle = {Proceedings of the Eleventh ACM International Conference on Future Energy Systems},
pages = {68–73},
numpages = {6},
keywords = {Federated Learning, Applied Machine Learning, HVAC Analytics},
location = {Virtual Event, Australia},
series = {e-Energy ’20}
}