BALTEEC - Bayesian Active Learning Techniques for Energy EfficienCy in buildings

 

Clemetine PrieurClémentine Prieur,
Full Professor at Grenoble Alpes University (UGA), Researcher at Jean Kuntzmann Laboratory (LJK) as a member of the Inria project-team AIRSEA, clementine.prieur@univ-grenoble-alpes.fr 


Rodolphe Le Riche

Rodolphe Le Riche,
CNRS senior researcher,
Laboratoire d’Informatique de Modélisation et d’Optimisation des Systèmes,
rodolphe.le_riche@uca.fr 










DESCRIPTION

Doing more with less: the main challenge of our project is to design new algorithms with the ability to adapt online to data structure and model complexity, with a special emphasis put on uncertainties. However, the research developments we propose are inspired by and applied to realistic issues of energy efficiency in buildings, with two advantages: first, the project has a clear environmental impact; second, it provides teaching and research applications to AI algorithms that can attract a public looking for clear positive societal benefits. The BALTEEC chair project is part of the MIAI Institute Axis “Core AI: Machine learning and reasoning”.

Work packages:

WP0: code for the test cases based on thermal simulations of buildings
WP1: active learning for conditional variables – application to the design and verification of energy renovations
WP2: dealing with uncertain time series – application to the control and fault diagnosis of energy in buildings equipped with a photovoltaic power generator
WP3: classification and multiscale models – application to the detection of energy renovation at the district scale

Innovation:

The BALTEEC project will primarily produce generic Bayesian learning procedures. The associated software prototypes will be available under an open source license. In addition, the test case platform gathering energy-in-buildings challenges will also be distributed under an open source license. Some of the algorithms developed, like those dealing with predictions of energy use, may be of interest to the three industrial partners as possible components of their commercial products. The chair intellectual property contract will keep this option possible yet with no exclusivity, to favour open science inside BALTEEC chair.

ACTIVITIES

The BALTEEC Chair is starting this 2025 fall semester:
a PhD student has been hired to work on WP1 from November 2025,
We hire : a postdoctoral position is open for one year along with another PhD position.

CHAIR PRESENTATION

Chair Présentation BALTEEC
License:
Unless otherwise stated, all documents are shared under the Creative Commons BY-NC-ND 4.0 license.
You may view and share them for non-commercial purposes, without modification, and with appropriate credit to the authors.
Published on  August 21, 2025
Updated on August 21, 2025