PCPL is a new framework for physical learning that extracts updates from measurable contrasts between perturbed states, unifying equilibrium and frequency propagation and demonstrated on spring networks and photonic circuits for classification and analog multiplication.
A deep learning theory for neu- ral networks grounded in physics.arXiv preprint arXiv:2103.09985, 2021
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Perturbative Contrastive Physical Learning
PCPL is a new framework for physical learning that extracts updates from measurable contrasts between perturbed states, unifying equilibrium and frequency propagation and demonstrated on spring networks and photonic circuits for classification and analog multiplication.