Spiking Neural Networks Hardware Implementations and Challenges: A Survey
Аппаратные реализации и проблемы реализации спайковых нейронных сетей
2019-04-05
SCID: 54.1/xjc734ze
Discuss with AI
event-driven algorithmshardware implementationsneuromorphic computingspiking neural networkstraining mechanisms
Figures from the paper
Abstract (AI)
Neuromorphic computing is henceforth a major research field for both academic and industrial actors. As opposed to Von Neumann machines, brain-inspired processors aim at bringing closer the memory and the computational elements to efficiently evaluate machine learning algorithms. Recently, spiking neural networks, a generation of cognitive algorithms employing computational primitives mimicking neuron and synapse operational principles, have become an important part of deep learning. They are expected to improve the computational performance and efficiency of neural networks, but they are best suited for hardware able to support their temporal dynamics. In this survey, we present the state of the art of hardware implementations of spiking neural networks and the current trends in algorithm elaboration from model selection to training mechanisms. The scope of existing solutions is extensive; we thus present the general framework and study on a case-by-case basis the relevant particularities. We describe the strategies employed to leverage the characteristics of these event-driven algorithms at the hardware level and discuss their related advantages and challenges.
Key Findings
1
Existing hardware solutions are diverse; the survey examines how they exploit event-driven characteristics and identifies corresponding advantages and implementation challenges.
2
Neuromorphic computing brings memory and computation closer together to improve the efficiency of machine-learning evaluation compared with von Neumann architectures.
3
Spiking neural networks are particularly suited to hardware architectures that support temporal dynamics and event-driven computation.
4
Spiking neural networks use neuron- and synapse-inspired computational primitives and are expected to improve neural-network computational performance and efficiency.
5
The survey reviews state-of-the-art spiking-neural-network hardware implementations and algorithmic trends spanning model selection through training mechanisms.
Research Object
hardware implementations of spiking neural networks
Research Subject
hardware-level strategies, advantages, challenges, and algorithmic design trends for efficiently supporting the temporal dynamics of spiking neural networks
Publication Details
Publication Date
2019-04-05
Journal
Publisher
ISSN
Cited by
343
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest