Dynamical stability conditions for recurrent neural networks with unsaturating piecewise linear transfer functions

Wersing H, Beyn W-J, Ritter H (2001)
Neural Computation 13(8): 1811-1825.

Journal Article | Published | English

No fulltext has been uploaded

Abstract
We establish two conditions that ensure the nondivergence of additive recurrent networks with unsaturating piecewise linear transfer functions, also called linear threshold or semilinear transfer functions. As Hahnloser, Sarpeshkar, Mahowald, Douglas, and Seung (2000) showed, networks of this type can be efficiently built in silicon and exhibit the coexistence of digital selection and analog amplification in a single circuit. To obtain this behavior, the network must be multistable and nondivergent, and our conditions allow determining the regimes where this can be achieved with maximal recurrent amplification. The first condition can be applied to nonsymmetric networks and has a simple interpretation of requiring that the strength of local inhibition match the sum over excitatory weights converging onto a neuron. The second condition is restricted to symmetric networks, but can also take into account the stabilizing effect of nonlocal inhibitory interactions. We demonstrate the application of the conditions on a simple example and the orientation-selectivity mo del of Ben-Yishai, Lev Bar-Or, and Sompolinsky (1995). We show that the conditions can be used to identify in their model regions of maximal orientation-selective amplification and symmetry breaking.
Publishing Year
ISSN
PUB-ID

Cite this

Wersing H, Beyn W-J, Ritter H. Dynamical stability conditions for recurrent neural networks with unsaturating piecewise linear transfer functions. Neural Computation. 2001;13(8):1811-1825.
Wersing, H., Beyn, W. - J., & Ritter, H. (2001). Dynamical stability conditions for recurrent neural networks with unsaturating piecewise linear transfer functions. Neural Computation, 13(8), 1811-1825.
Wersing, H., Beyn, W. - J., and Ritter, H. (2001). Dynamical stability conditions for recurrent neural networks with unsaturating piecewise linear transfer functions. Neural Computation 13, 1811-1825.
Wersing, H., Beyn, W.-J., & Ritter, H., 2001. Dynamical stability conditions for recurrent neural networks with unsaturating piecewise linear transfer functions. Neural Computation, 13(8), p 1811-1825.
H. Wersing, W.-J. Beyn, and H. Ritter, “Dynamical stability conditions for recurrent neural networks with unsaturating piecewise linear transfer functions”, Neural Computation, vol. 13, 2001, pp. 1811-1825.
Wersing, H., Beyn, W.-J., Ritter, H.: Dynamical stability conditions for recurrent neural networks with unsaturating piecewise linear transfer functions. Neural Computation. 13, 1811-1825 (2001).
Wersing, Heiko, Beyn, Wolf-Jürgen, and Ritter, Helge. “Dynamical stability conditions for recurrent neural networks with unsaturating piecewise linear transfer functions”. Neural Computation 13.8 (2001): 1811-1825.
This data publication is cited in the following publications:
This publication cites the following data publications:

25 Citations in Europe PMC

Data provided by Europe PubMed Central.

Scale-limited activating sets and multiperiodicity for threshold-linear networks on time scales.
Huang Z, Raffoul YN, Cheng CY., IEEE Trans Cybern 44(4), 2014
PMID: 23757562
Boundedness and complete stability of complex-valued neural networks with time delay.
Bo Zhou , Qiankun Song ., IEEE Trans Neural Netw Learn Syst 24(8), 2013
PMID: 24808563
Memory dynamics in attractor networks with saliency weights.
Tang H, Li H, Yan R., Neural Comput 22(7), 2010
PMID: 20235821
Nontrivial global attractors in 2-D multistable attractor neural networks.
Zou L, Tang H, Tan KC, Zhang W., IEEE Trans Neural Netw 20(11), 2009
PMID: 19884069
State-dependent computation using coupled recurrent networks.
Rutishauser U, Douglas RJ., Neural Comput 21(2), 2009
PMID: 19431267
Permitted and forbidden sets in discrete-time linear threshold recurrent neural networks.
Yi Z, Zhang L, Yu J, Tan KK., IEEE Trans Neural Netw 20(6), 2009
PMID: 19423436
Representations of continuous attractors of recurrent neural networks.
Yu J, Yi Z, Zhang L., IEEE Trans Neural Netw 20(2), 2009
PMID: 19150791
Analysis of continuous attractors for 2-D linear threshold neural networks.
Zou L, Tang H, Tan KC, Zhang W., IEEE Trans Neural Netw 20(1), 2009
PMID: 19129036
Selectivity and stability via dendritic nonlinearity.
Morita K, Okada M, Aihara K., Neural Comput 19(7), 2007
PMID: 17521280
Learning lateral interactions for feature binding and sensory segmentation from prototypic basis interactions.
Weng S, Wersing H, Steil JJ, Ritter H., IEEE Trans Neural Netw 17(4), 2006
PMID: 16856650
Dynamics analysis and analog associative memory of networks with LT neurons.
Tang H, Tan KC, Teoh EJ., IEEE Trans Neural Netw 17(2), 2006
PMID: 16566468
Analysis of cyclic dynamics for networks of linear threshold neurons.
Tang HJ, Tan KC, Zhang W., Neural Comput 17(1), 2005
PMID: 15563749
Discrimination networks for maximum selection.
Jain BJ, Wysotzki F., Neural Netw 17(1), 2004
PMID: 14690714
Analyzing stability of equilibrium points in neural networks: a general approach.
Truccolo WA, Rangarajan G, Chen Y, Ding M., Neural Netw 16(10), 2003
PMID: 14622876
Permitted and forbidden sets in symmetric threshold-linear networks.
Hahnloser RH, Seung HS, Slotine JJ., Neural Comput 15(3), 2003
PMID: 12620160
Selectively grouping neurons in recurrent networks of lateral inhibition.
Xie X, Hahnloser RH, Seung HS., Neural Comput 14(11), 2002
PMID: 12433293
Dynamic stability conditions for Lotka-Volterra recurrent neural networks with delays.
Yi Z, Tan KK., Phys Rev E Stat Nonlin Soft Matter Phys 66(1 Pt 1), 2002
PMID: 12241387
Attentional recruitment of inter-areal recurrent networks for selective gain control.
Hahnloser RH, Douglas RJ, Hepp K., Neural Comput 14(7), 2002
PMID: 12079551

18 References

Data provided by Europe PubMed Central.

A model for the intracortical origin of orientation preference and tuning in macaque striate cortex.
Adorjan P, Levitt JB, Lund JS, Obermayer K., Vis. Neurosci. 16(2), 1999
PMID: 10367965
Recurrent excitation in neocortical circuits.
Douglas RJ, Koch C, Mahowald M, Martin KA, Suarez HH., Science 269(5226), 1995
PMID: 7638624
Piecewise affine bijections of ?n, and the equation Sx+−Tx−=y
Kuhn, Linear Algebra and its Applications 96(1), 1987
A competitive-layer model for feature binding and sensory segmentation.
Wersing H, Steil JJ, Ritter H., Neural Comput 13(2), 2001
PMID: 11177439
On Neurodynamics with Limiter Function and Linsker's Developmental Model
Feng, Neural Computation 8(5), 1996

Lippmann, IEEE ASSP Magazine 4(2), 1987
Digital selection and analogue amplification coexist in a cortex-inspired silicon circuit.
Hahnloser RH, Sarpeshkar R, Mahowald MA, Douglas RJ, Seung HS., Nature 405(6789), 2000
PMID: 10879535
A model of multiplicative neural responses in parietal cortex.
Salinas E, Abbott LF., Proc. Natl. Acad. Sci. U.S.A. 93(21), 1996
PMID: 8876244

Feng, Journal of Physics A Mathematical and General 29(16), 1996
Convergent activation dynamics in continuous time networks
Hirsch, Neural Networks 2(5), 1989
Equilibria of the brain-state-in-a-box (BSB) neural model
Greenberg, Neural Networks 1(4), 1988
Self-organization of orientation sensitive cells in the striate cortex.
von der Malsburg C., Kybernetik 14(2), 1973
PMID: 4786750
Theory of orientation tuning in visual cortex.
Ben-Yishai R, Bar-Or RL, Sompolinsky H., Proc. Natl. Acad. Sci. U.S.A. 92(9), 1995
PMID: 7731993

Export

0 Marked Publications

Open Data PUB

Web of Science

View record in Web of Science®

Sources

PMID: 11506671
PubMed | Europe PMC

Search this title in

Google Scholar