On the Temperature of Machine Learning Systems

arXiv:2404.13218v1 Announce Type: new
Abstract: We develop a thermodynamic theory for machine learning (ML) systems. Similar to physical thermodynamic systems which are characterized by energy and entropy, ML systems possess these characteristics as well. This comparison inspire us to integrate the concept of temperature into ML systems grounded in the fundamental principles of thermodynamics, and establish a basic thermodynamic framework for machine learning systems with non-Boltzmann distributions. We introduce the concept of states within a ML system, identify two typical …

abstract arxiv comparison concept cs.ai cs.lg cs.ne energy entropy fundamental learning systems machine machine learning systems theory type