Limits...
Neuromorphic atomic switch networks.

Avizienis AV, Sillin HO, Martin-Olmos C, Shieh HH, Aono M, Stieg AZ, Gimzewski JK - PLoS ONE (2012)

Bottom Line: However, conventional fabrication techniques are unable to efficiently generate structures with the highly complex interconnectivity found in biological neuronal networks.Further, emergent behaviors unique to the complex network of atomic switches and akin to brain function are observed, namely spatially distributed memory, recurrent dynamics and the activation of feedforward subnetworks.These devices display the functional characteristics required for implementing unconventional, biologically and neurally inspired computational methodologies in a synthetic experimental system.

View Article: PubMed Central - PubMed

Affiliation: Department of Chemistry and Biochemistry, University of California Los Angeles, Los Angeles, California, United States of America.

ABSTRACT
Efforts to emulate the formidable information processing capabilities of the brain through neuromorphic engineering have been bolstered by recent progress in the fabrication of nonlinear, nanoscale circuit elements that exhibit synapse-like operational characteristics. However, conventional fabrication techniques are unable to efficiently generate structures with the highly complex interconnectivity found in biological neuronal networks. Here we demonstrate the physical realization of a self-assembled neuromorphic device which implements basic concepts of systems neuroscience through a hardware-based platform comprised of over a billion interconnected atomic-switch inorganic synapses embedded in a complex network of silver nanowires. Observations of network activation and passive harmonic generation demonstrate a collective response to input stimulus in agreement with recent theoretical predictions. Further, emergent behaviors unique to the complex network of atomic switches and akin to brain function are observed, namely spatially distributed memory, recurrent dynamics and the activation of feedforward subnetworks. These devices display the functional characteristics required for implementing unconventional, biologically and neurally inspired computational methodologies in a synthetic experimental system.

Show MeSH
DC Response – recurrent dynamics.(a) Time traces of current response to 1 V DC bias show large current increases and decreases at all time scales around a mean of 5.81 µA (172 kΩ); shorter time traces (ii–iii) are subsets of (i). Representative device parameters: ROFF>10 MΩ, RON<20 kΩ, VT = 3 V during activation (b) Fourier transforms of DC bias response for Ag control (grey) and functionalized Ag-Ag2S (black) networks. The power spectrum of the functionalized network displays 1/fβ power law scaling (β = 1.34).
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3412809&req=5

pone-0042772-g004: DC Response – recurrent dynamics.(a) Time traces of current response to 1 V DC bias show large current increases and decreases at all time scales around a mean of 5.81 µA (172 kΩ); shorter time traces (ii–iii) are subsets of (i). Representative device parameters: ROFF>10 MΩ, RON<20 kΩ, VT = 3 V during activation (b) Fourier transforms of DC bias response for Ag control (grey) and functionalized Ag-Ag2S (black) networks. The power spectrum of the functionalized network displays 1/fβ power law scaling (β = 1.34).

Mentions: Applying a constant 1 V DC bias (Figure 4a) produced persistent, bidirectional fluctuations—both increases and decreases—in network conductivity of large magnitudes (∼20–150%) over a range of time scales (seconds-hours). In the absence of recurrent structures within the network, the filamentary mechanism of an atomic switch implies that conductivity would increase monotonically under constant DC bias. The applied voltage leads to the thickening of filaments until the potential drop across the junctions is insufficient to reduce more silver cations [13]. However, large bidirectional fluctuations (ΔI greater than 100% on the scale of hours) in the current response persisted for several days under constant applied voltage, demonstrating that the complex network connectivity inherently resists localized positive feedback that would lead to the serial formation of a single, dominant high conductivity pathway between electrodes. Rather, recurrent loops in the network create complex couplings between switches, resulting in network dynamics that do not converge to a steady state even under constant bias. A single switch turning ON does not simply lead to an increased potential drop across the next junction in a serial chain, but redistributes voltage across many recurrent connections that can ultimately produce a net decrease in network conductivity. This behavior represents a network-scale analog of defect-defect interactions that have been observed to produce current fluctuations in metal nanobridges [54]. The nanoscale switch filaments couple these interactions with electrochemical redox processes, leading to significant changes in the conductivity state of the entire network.


Neuromorphic atomic switch networks.

Avizienis AV, Sillin HO, Martin-Olmos C, Shieh HH, Aono M, Stieg AZ, Gimzewski JK - PLoS ONE (2012)

DC Response – recurrent dynamics.(a) Time traces of current response to 1 V DC bias show large current increases and decreases at all time scales around a mean of 5.81 µA (172 kΩ); shorter time traces (ii–iii) are subsets of (i). Representative device parameters: ROFF>10 MΩ, RON<20 kΩ, VT = 3 V during activation (b) Fourier transforms of DC bias response for Ag control (grey) and functionalized Ag-Ag2S (black) networks. The power spectrum of the functionalized network displays 1/fβ power law scaling (β = 1.34).
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3412809&req=5

pone-0042772-g004: DC Response – recurrent dynamics.(a) Time traces of current response to 1 V DC bias show large current increases and decreases at all time scales around a mean of 5.81 µA (172 kΩ); shorter time traces (ii–iii) are subsets of (i). Representative device parameters: ROFF>10 MΩ, RON<20 kΩ, VT = 3 V during activation (b) Fourier transforms of DC bias response for Ag control (grey) and functionalized Ag-Ag2S (black) networks. The power spectrum of the functionalized network displays 1/fβ power law scaling (β = 1.34).
Mentions: Applying a constant 1 V DC bias (Figure 4a) produced persistent, bidirectional fluctuations—both increases and decreases—in network conductivity of large magnitudes (∼20–150%) over a range of time scales (seconds-hours). In the absence of recurrent structures within the network, the filamentary mechanism of an atomic switch implies that conductivity would increase monotonically under constant DC bias. The applied voltage leads to the thickening of filaments until the potential drop across the junctions is insufficient to reduce more silver cations [13]. However, large bidirectional fluctuations (ΔI greater than 100% on the scale of hours) in the current response persisted for several days under constant applied voltage, demonstrating that the complex network connectivity inherently resists localized positive feedback that would lead to the serial formation of a single, dominant high conductivity pathway between electrodes. Rather, recurrent loops in the network create complex couplings between switches, resulting in network dynamics that do not converge to a steady state even under constant bias. A single switch turning ON does not simply lead to an increased potential drop across the next junction in a serial chain, but redistributes voltage across many recurrent connections that can ultimately produce a net decrease in network conductivity. This behavior represents a network-scale analog of defect-defect interactions that have been observed to produce current fluctuations in metal nanobridges [54]. The nanoscale switch filaments couple these interactions with electrochemical redox processes, leading to significant changes in the conductivity state of the entire network.

Bottom Line: However, conventional fabrication techniques are unable to efficiently generate structures with the highly complex interconnectivity found in biological neuronal networks.Further, emergent behaviors unique to the complex network of atomic switches and akin to brain function are observed, namely spatially distributed memory, recurrent dynamics and the activation of feedforward subnetworks.These devices display the functional characteristics required for implementing unconventional, biologically and neurally inspired computational methodologies in a synthetic experimental system.

View Article: PubMed Central - PubMed

Affiliation: Department of Chemistry and Biochemistry, University of California Los Angeles, Los Angeles, California, United States of America.

ABSTRACT
Efforts to emulate the formidable information processing capabilities of the brain through neuromorphic engineering have been bolstered by recent progress in the fabrication of nonlinear, nanoscale circuit elements that exhibit synapse-like operational characteristics. However, conventional fabrication techniques are unable to efficiently generate structures with the highly complex interconnectivity found in biological neuronal networks. Here we demonstrate the physical realization of a self-assembled neuromorphic device which implements basic concepts of systems neuroscience through a hardware-based platform comprised of over a billion interconnected atomic-switch inorganic synapses embedded in a complex network of silver nanowires. Observations of network activation and passive harmonic generation demonstrate a collective response to input stimulus in agreement with recent theoretical predictions. Further, emergent behaviors unique to the complex network of atomic switches and akin to brain function are observed, namely spatially distributed memory, recurrent dynamics and the activation of feedforward subnetworks. These devices display the functional characteristics required for implementing unconventional, biologically and neurally inspired computational methodologies in a synthetic experimental system.

Show MeSH