@ClopathLab
www.pnas.org Learning predictive signals within a local recurrent circuit | PNAS The predictive coding hypothesis proposes that top–down predictions are compared with incoming bottom–up sensory information, with prediction error... 🚀Out in PNAS!View quoted post on X
We find that recurrent networks trained with the predictive plasticity rules explain many features of prediction errors observed in experimental studies
pnas.org/doi/10.1073/pn…
@ClopathLab
elifesciences.org Embedding stochastic dynamics of the environment in spontaneous activity by prediction-based... A computational model shows how synaptic plasticity shapes spontaneous activity to encode the transition statistics of sensory experience. Now inView quoted post on X
@eLife
“Biologically plausible excitatory/inhibitory plasticity rules in recurrent spiking networks embed state-transition statistics into spontaneous activity.”
doi.org/10.7554/eLife.…
@ClopathLab
arxiv.org Relational reasoning and inductive bias in transformers and large... Transformer-based models have demonstrated remarkable reasoning abilities, but the mechanisms underlying relational reasoning remain poorly understood. We investigate how transformers perform... 