Regenerative memory in time-delayed neuromorphic photonic resonators

Romeira, Bruno; Avó, R.; Figueiredo, J. M. L.; Barland, S.; Javaloyes, J.

doi:10.1038/srep19510

Cited by 123 publications

(101 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent experimental and theoretical results report that DS are also possible in systems with time-delayed feedback that do not include explicit spatial variables [18][19][20][21][22][23][24][25][26][27]. In these systems the time delay is larger than the other timescales and the DS are temporally localized.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Temporal Dissipative Solitons in Time-Delay Feedback Systems

et al. 2019

View full text Add to dashboard Cite

Localized states are a universal phenomenon observed in spatially distributed dissipative nonlinear systems. Known as dissipative solitons, auto-solitons, spot or pulse solutions, these states play an important role in data transmission using optical pulses, neural signal propagation, and other processes. While this phenomenon was thoroughly studied in spatially extended systems, temporally localized states are gaining attention only recently, driven primarily by applications from fiber or semiconductor lasers. Here we present a theory for temporal dissipative solitons (TDS) in systems with time-delayed feedback. In particular, we derive a system with an advanced argument, which determines the profile of the TDS. We also provide a complete classification of the spectrum of TDS into interface and pseudo-continuous spectrum. We illustrate our theory with two examples: a generic delayed phase oscillator, which is a reduced model for an injected laser with feedback, and the FitzHugh-Nagumo neuron with delayed feedback. Finally, we discuss possible destabilization mechanisms of TDS and show an example where the TDS delocalizes and its pseudo-continuous spectrum develops a modulational instability.

show abstract

mentioning

confidence: 99%

“…For instance, they can be composed of several pulses, known as soliton molecules or bound states [21,26,33]. Experimental and theoretical methods to control the nucleation or cancellation of TDS have been introduced in [22,23].…”

mentioning

confidence: 99%

Temporal Dissipative Solitons in Time-Delay Feedback Systems

et al. 2019

View full text Add to dashboard Cite

show abstract

“…The great majority of work on photonic neurons has focused on lasers that implement spiking models similar to biological neurons [43][44][45][46][47][48][49][50][51], reviewed in [52]. To claim a non-spiking modulator as a photonic neuron represents a departure from the origins of neuromorphic photonics.…”

Section: B Non-spiking Photonic Neuronsmentioning

confidence: 99%

“…cascadability). Research in neuromorphic photonics has seen an abundance of semiconductor lasers that exhibit nonlinear transfer functions, particularly those of a spiking neuron [43][44][45][46][47][48][49][50][51], reviewed in [52]. Nevertheless, all of the conditions of network-compatibility have yet to have been conclusively demonstrated in a single device (with exception of the fiber laser in [53]), and much of this research overlooks fan-in and/or cascadability entirely.…”

Section: Introductionmentioning

confidence: 99%

Silicon Photonic Modulator Neuron

et al. 2019

View full text Add to dashboard Cite

There has been a recently renewed interest in neuromorphic photonics, a field promising to access pivotal and unexplored regimes of machine intelligence. Progress has been made on isolated neurons and analog interconnects; nevertheless, this renewal has yet to produce a demonstration of a silicon photonic neuron capable of interacting with other like neurons. We report a modulator-class photonic neuron fabricated in a conventional silicon photonic process line. We demonstrate behaviors of transfer function configurability, fan-in, inhibition, time-resolved processing, and, crucially, autaptic cascadability -a sufficient set of behaviors for a device to act as a neuron participating in a network of like neurons. The silicon photonic modulator neuron constitutes the final piece needed to make photonic neural networks fully integrated on currently available silicon photonic platforms.In this work, we fabricate and demonstrate a silicon photonic modulator neuron. It consists of a balanced photodetector directly connected to a microring (MRR) modulator. We demonstrate that this device possesses the necessary capabilities of a network-compatible neuron: fan-in, high-gain optical-to-optical nonlinearity, and arXiv:1812.11898v1 [physics.app-ph]

show abstract

“…Hybrid systems, where for some or all layers self-feedback is removed, would incorporate feed-forward architectures [33]. Layers featuring excitable solitons can potentially create long term memory [34] and, when combined with the reported LP and BP-layers, physically implement long-short term memory networks [35]. This opens possibilities in new domains like natural language processing and sequence generation.…”

Section: Nodes Per Layermentioning

confidence: 99%

Coupled Nonlinear Delay Systems as Deep Convolutional Neural Networks

et al. 2019

View full text Add to dashboard Cite

Neural networks are transforming the field of computer algorithms, yet their emulation on current computing substrates is highly inefficient. Reservoir computing was successfully implemented on a large variety of substrates and gave new insight in overcoming this implementation bottleneck. Despite its success, the approach lags behind the state of the art in deep learning. We therefore extend time-delay reservoirs to deep networks and demonstrate that these conceptually correspond to deep convolutional neural networks. Convolution is intrinsically realized on a substrate level by generic drive-response properties of dynamical systems. The resulting novelty is avoiding vectormatrix products between layers, which cause low efficiency in today's substrates. Compared to singleton time-delay reservoirs, our deep network achieves accuracy improvements by at least an order of magnitude in Mackey-Glass and Lorenz timeseries prediction.

show abstract

Regenerative memory in time-delayed neuromorphic photonic resonators

Cited by 123 publications

References 58 publications

Temporal Dissipative Solitons in Time-Delay Feedback Systems

Temporal Dissipative Solitons in Time-Delay Feedback Systems

Silicon Photonic Modulator Neuron

Coupled Nonlinear Delay Systems as Deep Convolutional Neural Networks

Contact Info

Product

Resources

About