FW26, William Gates Building
\[\frac{\text{d}H}{\text{d}t} \geq 0\]
Maxwell’s Demon:
Clausius: the entropy of the universe tends to a maximum.
Same formula, three names: Boltzmann weights, Gibbs distribution, canonical ensemble.
The same energy-and-occupation grammar reappears in associative memory and early deep learning.
Once \(p_i = e^{-\beta E_i}/Z\) is fixed, thermodynamics is bookkeeping.
A thermodynamic bath is the large system that justifies the canonical ensemble: it fixes \(T\) and exchanges energy with a small system.
Equilibrium thermodynamics gives a prescription for reversible work. Real protocols take time.
Boltzmann machines (optional colour) of Ackley et al. (1985)
Hopfield networks (optional colour) of Hopfield (1982)
Why Stochastic Thermodynamics of Machine Learning? of Welling et al. (2026)
Chapter 3 of Welling et al. (2026)
Chapters 1–4 of Callen (1985)
Chapters 5–6 of Callen (1985)