Deep Learning With Spiking Neurons: Opportunities and Challenges

By Michael Pfeiffer, Thomas Pfeil. Frontiers in Neuroscience, 2018

750 citations recorded by the source corpus.

Spiking neural networks (SNNs) are inspired by information processing in biology, where sparse and asynchronous binary signals are communicated and processed in a massively parallel fashion. SNNs on neuromorphic hardware exhibit favorable properties such as low power consumption, fast inference, and event-driven information processing. This makes them interesting candidates for the efficient implementation of deep neural networks, the method of choice for many machine learning tasks. In this review, we address the opportunities that deep spiking networks offer and investigate in detail the challenges associated with training SNNs in a way that makes them competitive with conventional deep learning, but simultaneously allows for efficient mapping to hardware. A wide range of training methods for SNNs is presented, ranging from the conversion of conventional deep networks into SNNs, constrained training before conversion, spiking variants of backpropagation, and biologically motivated variants of STDP. The goal of our review is to define a categorization of SNN training methods, and summarize their advantages and drawbacks. We further discuss relationships between SNNs and binary networks, which are becoming popular for efficient digital hardware implementation. Neuromorphic hardware platforms have great potential to enable deep spiking networks in real-world applications. We compare the suitability of various neuromorphic systems that have been developed over the past years, and investigate potential use cases. Neuromorphic approaches and conventional machine learning should not be considered simply two solutions to the same classes of problems, instead it is possible to identify and exploit their task-specific advantages. Deep SNNs offer great opportunities to work with new types of event-based sensors, exploit temporal codes and local on-chip learning, and we have so far just scratched the surface of realizing these advantages in practical applications.

Subfields: Cellular and Molecular Neuroscience, Electrical and Electronic Engineering

View Paper: https://www.semanticscholar.org/paper/f7d8783648ef685eb020bda8a41b8bc5ced5c2a1

How researchers see this paper

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

In Cellular and Molecular Neuroscience

In Electrical and Electronic Engineering

Citations in the corpus

The Heidelberg Spiking Data Sets for the Systematic Evaluation of Spiking Neural Networks (2019, methodology): “…now promises to translate the success of gradient-based learning from conventional neural networks to the spiking domain [Neftci et al., 2019, Pfeiﬀer and Pfeil, 2018, Tavanaei et al., 2018] and thus to instantiate functional SNNs in-silico both on conventional computers and neuromorphic…”
Training Spiking Neural Networks Using Lessons From Deep Learning (2021, background): “A more rigorous treatment of shadow training techniques and challenges can be found in [162].”
Spiking-YOLO: Spiking Neural Network for Energy-Efficient Object Detection (2019, background): “Due to the event-driven nature, SNNs offer energy-efﬁcient operations (Pfeiffer and Pfeil 2018).”
A review of learning in biologically plausible spiking neural networks (2019, positive): “A population of neurons can be used instead of a single neuron for each class to reduce the variance of the output (Pfeiffer & Pfeil, 2018).”
Solving the Black Box Problem: A Normative Framework for Explainable Artificial Intelligence (2019, background): “For example, the artificial neural networks used for visual processing in self-driving cars are increasingly implemented on neuromorphic hardware devices that have considerable advantages with respect to speed and power consumption (Pfeiffer and Pfeil 2018).”
The remarkable robustness of surrogate gradient learning for instilling complex function in spiking neural networks (2020, methodology): “One approach is to use analog-valued machine learning data sets as input currents directly (Hunsberger & Eliasmith, 2015) or to first convert them to Poisson input spike trains (Lee et al., 2016; Pfeiffer & Pfeil, 2018).”
Memristive technologies for data storage, computation, encryption, and radio-frequency communication (2022, background): “With the incorporation of additional bioinspired neuronal and synaptic dynamics, SNNs could potentially outperform conventional deep learning in certain application domains, and memristive devices could be exploited to natively implement such dynamics ( 68 ).”