Meta learning complex plasticity rules with ES.

<p>(A) CMA-ES on I-to-E plasticity within the big polynomial search space for the stability task. Left: Evolution of the plasticity parameters and loss values along the optimization. Right: example network activity elicited by the meta learned I-to-E rule. (B) Same as A for I-to-E rules from t...

Full description

Saved in:

Bibliographic Details
Main Author:	Basile Confavreux (21175925) (author)
Other Authors:	Everton J. Agnes (13250463) (author), Friedemann Zenke (485069) (author), Henning Sprekeler (183138) (author), Tim P. Vogels (13250460) (author)
Published:	2025
Subjects:	Biophysics Neuroscience Sociology Infectious Diseases Biological Sciences not elsewhere classified Information Systems not elsewhere classified use evolutionary strategies local optimization strategy large spiking networks enable remain opaque various plasticity parameters devise loss functions long learning abilities identical network behaviour div >< p synaptic plasticity rules active plasticity rules synaptic plasticity plasticity rules active rules complex functions discover rules work provides using parameterizations search constraints priori </ previous work plausible solutions optimal rule mechanistic link makes exploration level computations key player increasing complexity guided discovery familiarity detection experimental limitations degeneracy cumbersome brain ’ balancing complexity
Tags:	Add Tag No Tags, Be the first to tag this record!

Meta learning complex plasticity rules with ES.

Similar Items