A typed accounting separates record correlation from operational capital value in finite learning devices, with separation, capitalization-efficiency, and value-retention theorems.
Stochastic Thermodynamics of Learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
Virtually every organism gathers information about its noisy environment and builds models from that data, mostly using neural networks. Here, we use stochastic thermodynamics to analyse the learning of a classification rule by a neural network. We show that the information acquired by the network is bounded by the thermodynamic cost of learning and introduce a learning efficiency $\eta\le1$. We discuss the conditions for optimal learning and analyse Hebbian learning in the thermodynamic limit.
fields
cond-mat.stat-mech 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Thermodynamics of Learning: A Typed Four-Component Accounting of Memory, Fit, and Value
A typed accounting separates record correlation from operational capital value in finite learning devices, with separation, capitalization-efficiency, and value-retention theorems.