A biologically plausible learning rule is claimed to always emerge to behave like generalized gradient descent

LZLiu Ziyin@LiuZiyin106:36 PM · Oct 17, 2026

In our recent PNAS work, we showed that a local, biologically plausible learning rule can be designed so that both the neural network and the learning rule itself are "learned." More importantly, this rule always emerges to behave like a generalized gradient descent.

arXiv.orgHow Biological Synapses Self-Assemble Gradient LearningExisting models of learning in the brain explain how given circuits learn, but not how biology could assemble those circuits in the first place. To address this gap, we formulate Self-Assembling...