[CTN] Seminar on Burstprop: a theory of learning Feb 22 at 14:30

[Britt Anderson]

Richard Naud of the University of Ottawa has wide interests in neural
coding (http://www.neurodynamic.uottawa.ca/neuralcoding/index.html) and
will be giving a UWloo Centre for Theoretical Neuroscience seminar this
coming Tuesday, Feb 22 at 14:30. The title and abstract for his talk
are:

Title: Burstprop: a mechanistic theory for the orchestration of learning
in cortex

Abstract: Synaptic plasticity is believed to be a key physiological
mechanism for learning. It is well established that it depends on pre-
and postsynaptic activity. However, models that rely solely on pre- and
postsynaptic activity for synaptic changes have, so far, not been able
to account for learning complex tasks that demand credit assignment in
hierarchical networks. Burstporp is a theory that exploit the fact that
synaptic plasticity is regulated by high-frequency bursts of spikes, it
allows pyramidal neurons higher in a hierarchical circuit to coordinate
the plasticity of lower-level connections. Using simulations and
mathematical analyses, we have demonstrated that, when paired with
short-term synaptic dynamics, regenerative activity in the apical
dendrites and synaptic plasticity in feedback pathways, a
burst-dependent learning rule can solve challenging tasks that require
deep network architectures. To end, I will provide a discussion on the
biological plausibility of burstprop as a theory for the coordination of
synaptic plasticity in the brain.

The meeting number and password for the zoom connection will follow
soon.

Thanks, Britt





-- 
Britt Anderson, PhD & MD
Director, Centre for Theoretical Neuroscience
Assoc. Prof., Dept. of Psychology
University of Waterloo, Canada
https://brittlab.uwaterloo.ca