English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Adaptive systems -- such as a biological organism gaining survival advantage,an autonomous robot executing a functional task, or a motor proteintransporting intracellular nutrients -- must model the regularities andstochasticity in their environments to take full advantage of thermodynamicresources. Analogously, but in a purely computational realm, machine learningalgorithms estimate models to capture predictable structure and identifyirrelevant noise in training data. This happens through optimization ofperformance metrics, such as model likelihood. If physically implemented, isthere a sense in which computational models estimated through machine learningare physically preferred? We introduce the thermodynamic principle that workproduction is the most relevant performance metric for an adaptive physicalagent and compare the results to the maximum-likelihood principle that guidesmachine learning. Within the class of physical agents that most efficientlyharvest energy from their environment, we demonstrate that an efficient agent'smodel explicitly determines its architecture and how much useful work itharvests from the environment. We then show that selecting the maximum-workagent for given environmental data corresponds to finding themaximum-likelihood model. This establishes an equivalence betweennonequilibrium thermodynamics and dynamic learning. In this way, workmaximization emerges as an organizing principle that underlies learning inadaptive thermodynamic systems.

Thermodynamic Machine Learning through Maximum Work Production