Unsupervised visual feature learning with spike-timing-dependent plasticity: How far are we from traditional feature learning approaches?

Abstract

Spiking neural networks (SNNs) equipped with latency coding and spike-timing dependent plasticity rules offer an alternative to solve the data and energy bottlenecks of standard computer vision approaches: they can learn visual features without supervision and can be implemented by ultra-low power hardware architectures. However, their performance in image classification has never been evaluated on recent image datasets. In this paper, we compare SNNs to autoencoders on three visual recognition datasets, and extend the use of SNNs to color images. The analysis of the results helps us identify some bottlenecks of SNNs: the limits of on-center/off-center coding, especially for color images, and the ineffectiveness of current inhibition mechanisms. These issues should be addressed to build effective SNNs for image recognition.

Multi-layered Spiking Neural Network with Target Timestamp Threshold Adaptation and STDP

“…We use a difference-of-Gaussians (DoG) filter to simulate on-off cells [10]. Without this pre-processing, SNNs fail to learn useful patterns, leading to low classification performances [14].…”

“…This mechanism is required to guarantee that neurons will learn different patterns, since only one neuron will apply STDP per sample. However, WTA inhibition drastically reduces the spiking activity, which can lead to poor classification performance [14]. For this reason, we remove the inhibition mechanism during the inference stage.…”

“…Instead of using numerical representations like traditional methods, SNNs use spikes to transmit information, which radically changes their learning mechanisms. However, their classification performances currently remain behind traditional deep learning methods [14]. This gap is in part due to the constraint of local computation in SNNs, which prevents the use of traditional training methods like back-propagation.…”

Brain-Inspired Self-Organization with Cellular Neuromorphic Computing for Multimodal Unsupervised Learning

“…We decided to use the second method with a CNN-based feature extraction as described in Reference [99]. We use supervised feature extraction to demonstrate that the ReSOM multimodal association is possible using features, then future works will focus on the transition to unsupervised feature extraction with commplex datasets based on the works of References [85,100]. Thus, we use a supervised CNN feature extractor with the LeNet-5 topology [101] except for the last convolution layer which has only 12 filters instead of 120.…”

EqSpike: Spike-driven equilibrium propagation for neuromorphic implementations

“…A large number of neuromorphic systems use the unsupervised and biologically inspired spike-timing-dependent plasticity (STDP) learning rule because its weight updates, based on the relative timing of pre- and post-synaptic spikes, are spatially local and can be achieved with compact circuits in several technologies ( Bi and Poo, 2001 ; Masquelier and Thorpe, 2007 ; Bichler et al, 2012 ; Zamarreño-Ramos et al, 2011 ; Jo et al, 2010 ; Pedretti et al, 2017 ; Serb et al, 2016 ; Prezioso et al, 2018 ; Thakur et al, 2018 ; Feldmann et al, 2019 ). Unfortunately, STDP weight updates generally do not minimize a global objective function for the network, and the accuracy of STDP-trained neural networks remains below state-of-the-art algorithms based on the error backpropagation ( Falez et al, 2019 ). Important research efforts therefore investigate how the error backpropagation algorithm can be mathematically modified to make it spatially local and appropriate for spiking neural networks ( Neftci et al, 2017 ; Sacramento et al, 2018 ; Richards et al, 2019 ; Neftci et al., 2019 ; Kaiser et al., 2020 ; Bellec et al, 2020 ; Payeur et al, 2020 ).…”