Structure optimization of reservoir networks

Roeschies, Benjamin; Igel, Christian

doi:10.1093/jigpal/jzp043

Cited by 17 publications

(16 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The spectral radius |λ max | of the weight matrix W plays a crucial role in determining the dynamics that will take place in the recurrent network. Other factors, such as smallworld degree, scale-free regimes, and bio-inspired axonal growth patterns, have also been shown to positively influence the capabilities of the reservoir [242]. On the other hand, a recent theoretical analysis by Zhang et al argues that all random reservoir topologies asymptotically converge to the same distribution of eigenvalues, implying that the topology is relatively indifferent after all [307].…”

Section: Reservoir Computingmentioning

confidence: 99%

Artificial Neurogenesis: An Introduction and Selective Review

Kowaliw

Bredèche

Chevallier³

et al. 2014

Growing Adaptive Machines

View full text Add to dashboard Cite

International audienceIn this introduction and review—like in the book which follows—we explore the hypothesis that adaptive growth is a means of producing brain-like machines. The emulation of neural development can incorporate desirable characteristics of natural neural systems into engineered designs. The introduction begins with a review of neural development and neural models. Next, artificial development— the use of a developmentally-inspired stage in engineering design—is introduced. Several strategies for performing this " meta-design " for artificial neural systems are reviewed. This work is divided into three main categories: bio-inspired representations ; developmental systems; and epigenetic simulations. Several specific network biases and their benefits to neural network design are identified in these contexts. In particular, several recent studies show a strong synergy, sometimes interchange-ability, between developmental and epigenetic processes—a topic that has remained largely under-explored in the literature

show abstract

Section: Reservoir Computingmentioning

confidence: 99%

Artificial Neurogenesis: An Introduction and Selective Review

Kowaliw

Bredèche

Chevallier³

et al. 2014

Growing Adaptive Machines

View full text Add to dashboard Cite

show abstract

“…To this end, there are various methods in the literature (Chatzidimitriou and Mitkas 2013; Roeschies and Igel 2010;Whiteson and Stone 2006). In Whiteson and Stone (2006), NEAT is used to develop ad hoc neural networks without recurrent connections so that learning can be applied using standard error backpropagation updates to further adapt weights toward a solution.…”

Section: Evolutionary Neural Networkmentioning

confidence: 99%

“…In Whiteson and Stone (2006), NEAT is used to develop ad hoc neural networks without recurrent connections so that learning can be applied using standard error backpropagation updates to further adapt weights toward a solution. In Roeschies and Igel (2010), the authors use their own evolutionary method to develop ESNs with competitive results in time series prediction problems. In Chatzidimitriou and Mitkas (2013), reservoirs start minimally and grow through evolution to solve the problem at hand.…”

Section: Evolutionary Neural Networkmentioning

confidence: 99%

Effective connectivity at synaptic level in humans: a review and future prospects

Gürcan

2014

Biol Cybern

View full text Add to dashboard Cite

Correct knowledge of the effective connectivity at the synaptic level in humans is a key prerequisite for increasing our understanding of the operation of the human central nervous system. Unfortunately, none of the current ambitious collaborative neuroscience projects pay enough attention to this topic and are thus unable to completely relate the microlevel properties of the system to its emergent macrolevel behaviors. In this review article, the problem of effective connectivity at the synaptic level in humans is explained, existing and possible computational approaches to fill explanatory gaps are reviewed, and the requisite characteristics of these approaches are considered.

show abstract

“…However, to create networks that will adapt to the problem at hand, autonomously, without knowing the capacity (in terms of neurons) required to handle the complexity of the problem, we need evolutionary function approximation methods. To this end, there are various methods in the literature (Chatzidimitriou and Mitkas 2013;Roeschies and Igel 2010;Whiteson and Stone 2006). Basically, these methods select automatically function approximator representations that enable efficient individual learning.…”

Section: Evolutionary Neural Networkmentioning

confidence: 99%

Mimicking human neuronal pathways in silico: an emergent model on the effective connectivity

Gürcan

Türker

Mano³

et al. 2013

J Comput Neurosci

View full text Add to dashboard Cite

We present a novel computational model that detects temporal configurations of a given human neuronal pathway and constructs its artificial replication. This poses a great challenge since direct recordings from individual neurons are impossible in the human central nervous system and therefore the underlying neuronal pathway has to be considered as a black box. For tackling this challenge, we used a branch of complex systems modeling called artificial self-organization in which large sets of software entities interacting locally give rise to bottom-up collective behaviors. The result is an emergent model where each software entity represents an integrate-and-fire neuron. We then applied the model to the reflex responses of single motor units obtained from conscious human subjects. Experimental results show that the model recovers functionality of real human neuronal pathways by comparing it to appropriate surrogate data. What makes the model promising is the fact that, to the best of our knowledge, it is the first realistic model to self-wire an artificial neuronal network by efficiently combining neuroscience with artificial self-organization. Although there is no evidence yet of the model's connectivity mapping onto the human connectivity, we anticipate this model will help neuroscientists to learn much more about human neuronal networks, and could also be used for predicting hypotheses to lead future experiments.

show abstract

Structure optimization of reservoir networks

Cited by 17 publications

References 24 publications

Artificial Neurogenesis: An Introduction and Selective Review

Artificial Neurogenesis: An Introduction and Selective Review

Effective connectivity at synaptic level in humans: a review and future prospects

Mimicking human neuronal pathways in silico: an emergent model on the effective connectivity

Contact Info

Product

Resources

About