On the computational power of circuits of spiking neurons

By Wolfgang Maass, Henry Markram. Journal of computer and system sciences (Print), 2004

327 citations recorded by the source corpus.

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

View Paper: https://www.semanticscholar.org/paper/0dce4e8f14052e6569ca0b1b4d012e24d938b0f6

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

Training Deep Spiking Neural Networks Using Backpropagation (2016, background): “s to gain efﬁciency, especially if they are combined with inputs from event-based sensors, which reduce redundant information based on asynchronous event processing [2, 19, 22]. Even though in theory [17] SNNs have been shown to be as computationally powerful as conventional artiﬁcial neural network”
Making Sense of Spatio-Temporal Preserving Representations for EEG-Based Human Intention Recognition (2020, background): “Although in theory the SNNs have shown to be as computationally powerful as conventional ANN [52], practically SNNs have not quite reached the same accuracy levels of ANNs in traditional machine learning tasks [53].”
Recent Advances in Physical Reservoir Computing: A Review (2018, methodology): “The LSM was proposed by Maass et al. (Maass, 2011; Maass et al., 2002) to explore the computational capability of neural microcircuits in the brain (Maass & Markram, 2004).”
Machine Learning for Wireless Networks with Artificial Intelligence: A Tutorial on Neural Networks (2017, background): “powerful [145], a major issue is the lack of adequate training algorithms suitable for deep SNNs.”
Weighted Sums of Random Kitchen Sinks: Replacing minimization with randomization in learning (2008, background): “But recently, architectures that randomly transform their inputs have been resurfacing in the machine learning community [2, 3, 4, 5], largely motivated by the fact that randomization is computationally cheaper than optimization.”
Artificial Neural Networks-Based Machine Learning for Wireless Networks: A Tutorial (2017, background): “n fact, deep SNNs are attractive for real-time applications where speed and power consumption are important factors. Although deep SNNs have been shown theoretically to be computationally 44 powerful [145], a major issue is the lack of adequate training algorithms suitable for deep SNNs. This follow”
From modulated Hebbian plasticity to simple behavior learning through noise and weight saturation (2012, background): “An interesting research direction in this context is the combination of the proposed learning structure with reservoir recurrent networks (Maass and Markram, 2004) to complement nonlinear computation with modulationdriven learning.”
Uniform approximation of functions with random bases (2008, background): “Recently, largely motivated by the fact that randomization is computationally cheaper than optimization, architectures based on random nonlinearities have been regaining popularity in the machine learning community for large scale data analysis [2], [3], [4], [5].”