method · first seen 5 Oct 2026
equilibrium propagation
Also: EqProp
A training method for energy-based networks that estimates gradients from the difference between two relaxed equilibrium states, avoiding backpropagation.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Lock-in EP: An In-Situ Training Algorithm for Oscillatory HardwarePapers · 5 Oct 2026 · Sowjanya Tammali, Wilkie Olin-Ammentorp
- An all-transistor physical-learning transformer can be trained to perform in-context learningPapers · 8 Oct 2026 · Michael C. Abbott, Benjamin B. Machta
- PotentialEP: Backprop-Free Training of CIFAR-Scale Generative Models via Equilibrium PropagationPapers · 10 Oct 2026 · Alex Gower
- Generative Replay for Equilibrium Propagation: Toward Edge-Friendly Continual Learning in Hopfield NetworksPapers · 27 Oct 2026 · Victor Cook, Gilles Sassatelli, Thierry Gil, Mohamed Watfa, Marina Reyboz
Latest sources
- Generative Replay for Equilibrium Propagation: Toward Edge-Friendly Continual Learning in Hopfield NetworksPapers · 27 Oct 2026 · Victor Cook, Gilles Sassatelli, Thierry Gil, Mohamed Watfa, Marina Reyboz
- PotentialEP: Backprop-Free Training of CIFAR-Scale Generative Models via Equilibrium PropagationPapers · 10 Oct 2026 · Alex Gower
- An all-transistor physical-learning transformer can be trained to perform in-context learningPapers · 8 Oct 2026 · Michael C. Abbott, Benjamin B. Machta
- Lock-in EP: An In-Situ Training Algorithm for Oscillatory HardwarePapers · 5 Oct 2026 · Sowjanya Tammali, Wilkie Olin-Ammentorp