M37.5 CONNECT THE MECHANISM
Learn with spike timing or an approximate backward signal
Fire 10 ms early and a synapse grows; fire 10 ms late and it shrinks. Meet the brain's timing rule, then the trick that lets backpropagation train spiking networks anyway.
LESSON OVERVIEW14 min lesson
Lesson overview
Fire 10 ms early and a synapse grows; fire 10 ms late and it shrinks. Meet the brain's timing rule, then the trick that lets backpropagation train spiking networks anyway.
What you’ll explore
- STDP changes synapses using relative spike timing, while surrogate-gradient methods replace an unhelpful spike derivative during training; these are distinct learning mechanisms with different biological and optimization assumptions.
GO TO THE SOURCE
Original explanations, connected to the research.
Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type (Bi & Poo, 1998)Surrogate Gradient Learning in Spiking Neural Networks (Neftci, Mostafa & Zenke, 2019)Neuronal Dynamics, chapter 19: synaptic plasticity and learning (Gerstner, Kistler, Naud & Paninski, 2014)Suggest a correction
A precise note can make an explanation better.
Choose the scene and describe what needs attention. Download a feedback file to share through a channel you already use. This page does not send feedback or connect you with a reviewer.