Abstract
We propose a solution to the weight transport problem, which questions the biological plausibility of the backpropagation algorithm. We derive our method based upon a theoretical analysis of the (approximate) dynamics of leaky integrate-and-fire neurons. Our results demonstrate that the use of spike timing alone outcompetes existing biologically plausible methods for synaptic weight inference in spiking neural network models. Furthermore, our proposed method is more flexible, being applicable to any spiking neuron model, is conservative in how many parameters are required for implementation and can be deployed in an online-fashion with minimal computational overhead. These features, together with its biological plausibility, make it an attractive mechanism for weight inference at single synapses.
Highlights
Backpropagation of error is a successful approach for training rate-based neural network models [1, 2]
We show that under a number of conditions our method outperforms both the Akrout and Regression Discontinuity Design (RDD) methods when applied to weight estimation in spiking neural network models
To verify the validity of our proposed spike-timing-dependent weight inference (STDWI) rule and demonstrate its flexibility, we compare it against a Bayesoptimal method for inferring synaptic inputs to a neuron with internal state modelled by a Wiener process (Figure 3)
Summary
Backpropagation of error is a successful approach for training rate-based neural network models [1, 2]. A number of attempts have been made to explain mechanisms by which backpropagation’s implausibilities can be addressed These can be divided into methods which propose alternative implementations of backpropagation, namely energy-based and dynamical systems methods which converge to backpropagation of error [10, 11, 12], for an overview see [6], and methods which show that components which are considered implausible can be approximated using alternative and plausible computations [13, 14, 15, 16, 17].
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
