Overcoming the Weight Transport Problem via Spike-Timing-Dependent Weight Inference
This addresses a key challenge in making neural networks more biologically realistic, though it is incremental as it builds on prior biologically plausible methods.
The paper tackled the weight transport problem in biologically plausible backpropagation by proposing a spike-timing-dependent method for synaptic weight inference in spiking neural networks, showing it outperforms existing methods in accuracy and flexibility.
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. We show 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 underlying weight inference at single synapses.