analiza aiqbench
This research introduces a novel metric within Bayesian optimization frameworks to explicitly balance predictive accuracy with computational energy consumption. By guiding parameter settings toward "frugal" configurations, the study demonstrates that optimizing for lower energy footprints can simultaneously improve convergence on complex design problems. The work addresses a critical sustainability challenge in machine learning by integrating environmental constraints directly into the optimization loop rather than treating them as an afterthought.
Energy-aware frugal Bayesian optimization
arXiv:2609.31638v1 Announce Type: new Abstract: Modern design optimization frameworks aim first and foremost for models with the most accurate predictions without balancing computational overhead. It remains a reason why scaled architecture and multidisciplinary design optimization problems are difficult to address, even with sample-efficient Bayesian optimizers. In this paper, a metric quantifying the computational energy footprint is introduced within a Bayesian optimization framework to guide the parameter setting of a model towards configurations that balance both performance and frugality. The computer experiments highlighted existing tradeoffs between optimum convergence and the underlying energy footprint, and sometimes resulted in both a better-found optimum and lower energy consumption.
Abstract:Modern design optimization frameworks aim first and foremost for models with the most accurate predictions without balancing computational overhead. It remains a reason why scaled architecture and multidisciplinary design optimization problems are difficult to address, even with sample-efficient Bayesian optimizers. In this paper, a metric quantifying the computational energy footprint is introduced within a Bayesian optimization framework to guide the parameter setting of a model towards configurations that balance both performance and frugality. The computer experiments highlighted existing tradeoffs between optimum convergence and the underlying energy footprint, and sometimes resulted in both a better-found optimum and lower energy consumption.
Published in MATEC Web of Conferences 422, 02009. 7th International Conference on Engineering Optimization
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Optimization and Control (math.OC); Classical Physics (physics.class-ph)
arXiv:2609.31638 [cs.LG]
(or arXiv:2609.31638v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2609.31638 Focus to learn more arXiv-issued DOI via DataCite
G. Plat, P. Saves, N. Bartoli, T. Lefebvre, J. Morlier, Energy-aware frugal Bayesian optimization, EngOpt, 2026, Volume 422, 2026, pp 02009
https://doi.org/10.1051/matecconf/202642202009 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 246 cuvinte
Subiecte conexe
Frugal Bayesian Optimization: Scalable Surrogates for Data- and Resource-Limited Discovery
arXiv · 3 aug. 2026
Enhanced measurements on quantum computers via the simultaneous probing of non-commuting Pauli operators
Quantum Journal · 7 oct. 2026
Curvature-aware Expected Free Energy as an Acquisition Function for Bayesian Optimization
arXiv · 17 sept. 2026
Human-in-the-Loop Meta Bayesian Optimization for Fusion Energy and Scientific Applications
arXiv · 16 sept. 2026