Research Salon

Accurate online training of dynamical spiking neural networks through Forward Propagation Through Time

https://arxiv.org/abs/2112.11231

Background

  • Training Recurrent Neural Networks via Forward Propagation Through Time
    https://icml.cc/virtual/2021/spotlight/10384
     
  • Liquid Time-constant Networks https://arxiv.org/abs/2006.04439
     
  • SNNs have demonstrated competitive and energy-efficient applications so far, e.g. Yin et al. 2021

Backprop Through Time (BPTT)

  • Training instabilities: vanishing or exploding gradients

     
  • Memory overhead Ω(T)
  • cannot be applied online: network is 'locked' during backward pass

Forward Propagation Through Time (FPTT)

  • FPTT takes a gradient step to minimize an instantaneous risk function at each time step

  • straightforward application of FPTT to SNNs fails

FPTT objective functions

Intermediate Losses for Terminal Prediction

with

Pseudo code

Liquid Time Constant SNN

Hasani et al., 2020

--> time constants are a function of inputs and hidden states

The Add task

One data point consists of two sequences (x1, x2) of length T and a target label y.

Example:

x1: <0,    0,    0,    1,     0,    0,    1,    0>

x2: <0.1, 0.6, 0.7, 0.2, 0.5, 0.8, 0.4, 0.3>

y:   <0,     0,    0,    0.2, 0.2, 0.2, 0.6, 0.6>

ASRNN: Adaptive Spiking RNN

LTC-SRNN: Liquid Time Constant Spiking RNN

The Add task

DVS Gesture

  • 20 - 500 frames are pre-processed using an SCNN
  • pre-processed input is flattened to 1D inputs

FPTT improves over online approximations of BPTT

Memory requirements and training time

Conclusions

  • FPTT works very well for long sequence learning with recurrent SNNs as long as LTC neurons are used.
     
  • It's not slower (with room for improvement) and has constant memory requirements.
     
  • Potentially very interesting in an online setting on neuromorphic hardware

Open Questions

  • How stable is training really?
     
  • Could this be applied to non-recurrent architectures?
     
  • Could this be combined with self-supervised learning?

FPTT research salon

By Gregor Lenz

FPTT research salon

  • 157