Learning Real-World Stimuli in a Neural Network with Spike-Driven Synaptic Dynamics

By Joseph M. Brader, Walter Senn, Stefano Fusi. Neural Computation, 2007

367 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Cognitive Neuroscience, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/898c54f4896e41a8885d465fc67925e159d4113d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Cognitive Neuroscience

In Electrical and Electronic Engineering

Citations in the corpus

STDP-based spiking deep neural networks for object recognition (2016, background): “The second group consists of SNNs with bioinspired learning rules which have shallow architectures [3, 17, 44, 59, 11] or only one trainable layer [36, 2, 23].”
Computational models of long term plasticity and memory (2017, methodology): “Synaptic models that are biologically plausible implementations of the perceptron algorithms have been proposed in the last decades (see e.g.[Brader et al., 2007, Legenstein et al., 2005a]).”
Deep Learning in Spiking Neural Networks (2018, background): “Spiking networks consisting of many LIF neurons equipped by spike-based synaptic plasticity rules have shown success in different pattern recognition tasks [128], [129].”
Training Spiking Neural Networks Using Lessons From Deep Learning (2021, background): “In other words, weight updates from BPTT closely resemble weight updates from STDP learning curves (see Appendix C2) [33].”
Unsupervised learning of digit recognition using spike-timing-dependent plasticity (2015, methodology): “Other spike-based learning methods often rely on diﬀerent variants of models of STDP (Brader et al., 2007; Habenschuss et al., 2012; Beyeler et al., 2013; Querlioz et al., 2013; Zhao et al., 2014), providing a closer match to biology for the learning procedure.”
Memory and Information Processing in Neuromorphic Systems (2015, methodology): “Note that although the analog circuits in the device have a high degree of variability (with a coefficient of variation of about 10% [134]), and that the learning process is stochastic [135], the attractors are formed reliably, and their population response is sustained robustly.”
Spiking Deep Networks with LIF Neurons (2015, methodology): “…ﬁring rate across all neurons in the network was 148 spikes/s, estimated from 20 test examples (30 seconds of Source MNIST error Brader et al. [8] 3.5% (1.3% misclassiﬁed, 2.2% not classiﬁed) (IF) Eliasmith et al. [9] 6% (LIF) Neftci et al. [10] 8.1% (LIF) O’Connor et al. [11] 2.52%…”
Sleep and synaptic renormalization: a computational study. (2010, methodology): “In particular, NMDARs are fundamental in triggering 296 AMPAR-mediated LTP and LTD (Carroll et al. 2001; Lu et al. 2001; Castellani et al. 2001; 297 Collingridge et al. 2004; Brader et al. 2007), while GABARs are involved in the modulation of 298 NMDARs activity (Bliss & Collinridge 1993).”
The Convallis Rule for Unsupervised Learning in Cortical Networks (2013, background): “F ’ ( V ( t )) P N s i ~ 1 K ( t { t si ) e { ( t { t ) = T d t , and then passed through a nonlinear shrinkage function H to ensure plasticity only occurs in response to multiple coincidences: ~ H ( Y ( t )) ([22,24,35]; see Materials and Methods for more details).”
Dynamic evolving spiking neural networks for on-line spatio- and spectro-temporal pattern recognition. (2013, methodology): “(2009) or in Brader et al. (2007)).”
Enabling Spike-Based Backpropagation for Training Deep Neural Network Architectures (2019, methodology): “The popular methods for SNN training are “Spike Time Dependent Plasticity (STDP)” based unsupervised learning (Brader et al., 2007; Diehl and Cook, 2015; Srinivasan et al., 2018a,b) and “spike-based backpropagation” based supervised learning (Lee et al., 2016; Neftci et al., 2017; Jin et al., 2018;…”
A 0.086-mm$^2$ 12.7-pJ/SOP 64k-Synapse 256-Neuron Online-Learning Digital Spiking Neuromorphic Processor in 28-nm CMOS (2018, background): “As the biological fan-in is on the order of 100 to 10,000 synapses per neuron, embedding locally an online learning rule such as spike-timing-dependent plasticity (STDP) [32] or spikedriven synaptic plasticity (SDSP) [33] in each single synapse is challenging [34].”
Minerva: Enabling Low-Power, Highly-Accurate Deep Neural Network Accelerators (2016, background): “Spiking Neural Networks (SNNs) [57, 58, 59, 60] are popular in ASIC designs, largely due to their compact mapping onto fully-parallel hardware.”
Eligibility Traces and Plasticity on Behavioral Time Scales: Experimental Support of NeoHebbian Three-Factor Learning Rules (2018, background): “For example, in some voltage-based (Brader et al., 2007) or calcium-based models (Shouval et al.”
Synaptic Consolidation: From Synapses to Behavioral Modeling (2015, background): “This integrate-and-threshold mechanism secures that ongoing small weight fluctuations cannot lead to long-term changes (Brader et al., 2007; Elliott and Lagogiannis, 2012).”
Neuromorphic Electronic Circuits for Building Autonomous Cognitive Systems (2014, methodology): “This is the condition for linearity, under which the circuit starts to behave as a first order low-pass filter, as described in Section II.”