pith. sign in

arxiv: 2604.22620 · v1 · submitted 2026-04-24 · ⚛️ physics.optics

Memory in Integrated Photonic Neural Networks: From Physical Mechanisms to Neuromorphic Architectures

Alessandro Foradori , Ilya Auslender , Stefano Biasi , Stefano Gretter , Alessio Lugnan , Emiliano Staffoli , Lorenzo Pavesi This is my paper

Pith reviewed 2026-05-08 10:29 UTC · model grok-4.3

classification ⚛️ physics.optics
keywords neuromorphic photonicsmemory mechanismssilicon photonicsreservoir computingphase-change materialsoptical neural networksintegrated opticsspiking networks
0
0 comments X

The pith

Physical mechanisms such as delay lines, slow-light effects, and phase-change materials supply memory in silicon photonics, allowing neuromorphic architectures to co-locate computation and storage.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The review classifies volatile and non-volatile memory mechanisms available in silicon-on-insulator and hybrid platforms, ranging from delay lines and multistable dynamics to charge trapping and phase-change materials. It connects these mechanisms to neural network topologies including feed-forward, reservoir computing, spiking, and hybrid recurrent systems, showing their use for time-dependent tasks like channel equalization. A sympathetic reader would care because the separation of memory and processing in conventional computers creates a growing bottleneck as networks scale, whereas photonic substrates can embed both functions in the same physical layer through light propagation and material properties. The authors draw from digital, biological, and photonic precedents to outline a classification that supports scalable, energy-efficient hardware.

Core claim

Memory in integrated photonic neural networks arises from intrinsic physical processes in silicon platforms that simultaneously store and process information, and these processes can be grouped by mechanism to guide the construction of neuromorphic systems that emulate biological co-localization of memory and computation.

What carries the argument

Classification of memory mechanisms according to their physical principles, from delay lines and slow-light structures for short-term effects to multistable dynamics, charge trapping, and phase-change materials for long-term storage, which then map onto specific neural architectures.

If this is right

  • Reservoir computing and spiking photonic networks become viable for real-time signal processing tasks such as telecommunications equalization.
  • Hybrid optoelectronic recurrent systems gain concrete design options from the separation of volatile and non-volatile memory types.
  • Energy efficiency gains follow directly when memory and processing share the same optical path rather than requiring separate data movement.
  • The classification identifies specific challenges in material stability and integration that must be solved for hardware deployment.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The framework could be used to predict which memory type matches a given task timescale before fabrication.
  • Combining photonic memory elements with electronic control layers may accelerate near-term prototypes while retaining optical bandwidth advantages.
  • Testing the classification on a benchmark like real-time speech recognition would reveal whether the physical mechanisms deliver measurable speed or power benefits.

Load-bearing premise

The reviewed physical mechanisms can be integrated into complete architectures at scale without prohibitive optical losses or fabrication complexity.

What would settle it

A working large-scale photonic neuromorphic chip built from the classified mechanisms that consumes more energy per operation or achieves lower accuracy than an electronic baseline on a standardized time-series task would falsify the practical value of the framework.

Figures

Figures reproduced from arXiv: 2604.22620 by Alessandro Foradori, Alessio Lugnan, Emiliano Staffoli, Ilya Auslender, Lorenzo Pavesi, Stefano Biasi, Stefano Gretter.

Figure 1
Figure 1. Figure 1: Conceptual comparison of structural (non-volatile) and dynamic (volatile) memory across digital, biological, and in￾tegrated photonic systems. Structural memory corresponds to long-lived parameters (e.g. weights, device states), while dynamic memory is encoded in transient state variables that evolve during computation. 1.7 The Physics of Memory and Learning Learning in any adaptive system requires that in… view at source ↗
Figure 2
Figure 2. Figure 2: The general description of biological memory from short-term (STM) to long-term memory (LTM), through encoding, learning and retrieval of information. Reproduced from [80], licensed under CC BY 4.0. 2.1.2 Biological Foundations of Memory: From Synaptic Communication to Engram Consolidation The formation and stabilization of memories emerge from a hierarchy of cellular, synaptic, and systems-level processes… view at source ↗
Figure 3
Figure 3. Figure 3: A general scheme of a biological neuron and a synapse with their main components. The terminal of a presynaptic neuron comes into close contact with a postsynaptic cell at the synapse. Adapted from SwissBioPics, licensed under CC BY 4.0 (indications of cell components are added) synaptic transmission. Phosphorylation facilitates the trafficking and membrane insertion of additional AMPA recep￾tors into the … view at source ↗
Figure 4
Figure 4. Figure 4: Synaptic communication and LTP. An action potential on the pre- synaptic cell stimulates the release of the neuro￾transmitters into the synaptic cleft by the synaptic vesicles. In the case of long term potentiation, the involved neurotransmitter is glutamate (1). It can bind to AMPA and NMDA receptors on the postsynaptic membrane. (2) The binding with glutamate induces AMPA receptor opening and a flux of N… view at source ↗
Figure 5
Figure 5. Figure 5: Cellular and synaptic organization across memory stages, from encoding to the formation of engram cells. The process begins with an increase in intrinsic neuronal excitability during encoding, which biases certain neurons toward allocation into the engram. During learning, activity in these cells promotes synaptic consolidation, leading to strengthened connections that can be further reinforced upon reacti… view at source ↗
Figure 6
Figure 6. Figure 6: (a) The synaptic model integrates the interplay between various mechanisms of calcium-dependent synaptic plasticity and the STC hypothesis. (b) Schematic of a part of the neural network that consists of excitatory (blue and dark blue circles) and inhibitory neurons (red circles), and receives external input from other brain areas (green arrows). Only synapses between excitatory neurons (blue and dark blue … view at source ↗
Figure 7
Figure 7. Figure 7: Schematic overview of the proposed classification of integrated photonic memory in silicon photonics. Memory mech￾anisms are categorized as volatile or non-volatile, with volatile memory further subdivided into response-induced and multistable￾induced mechanisms according to the physical process responsible for memory formation. Non-volatile memory is further classi￾fied, based on the same principle, into … view at source ↗
Figure 8
Figure 8. Figure 8: Schematic representation of three used architectures for implementing time delays in PICs. (a) Parallel, (b) serial and (c) recirculating-loop optical delay line scheme in (1) bar state and (2) cross state. Optical switching in an integrated photonic circuit is typically realized using an interferometric configuration incorporating MMI couplers and phase shifters, enabling active control of the different o… view at source ↗
Figure 9
Figure 9. Figure 9: Structures achieving slow light by means of (a) a MRR coupled to a Bragg grating cavity [138], (b) two coupled MRRs in a Side-Coupled Integrated Spaced Sequence of Resonators (SCISSOR) configuration where the yellow box shows the narrow transparency peak which appears within the broad coupled MRR resonance [140], (c) two coupled photonic crystal cavities [139], (d) a series of coupled-resonator optical wav… view at source ↗
Figure 10
Figure 10. Figure 10: Nonlinear effects and fading memory in a silicon MRR. (a) Two-photon absorption (TPA) of the cavity modes leads to the generation of free carriers (∆N), as illustrated by the energy–momentum diagram for indirect band-gap silicon. The relaxation of these carriers results in an increase in the effective cavity temperature (∆T). Both ∆N and ∆T feed back into the system, modifying the cavity modes. (b) Shift … view at source ↗
Figure 11
Figure 11. Figure 11: Phase-space dynamics, basins of attraction, and bifurcation-induced transitions in two-state nonlinear systems. The system is described by two variables x1 and x2. (a), (d) Bi-dimensional phase space containing attractors (red), repellers (blue), and trajectories (black). Basins of attraction are shown in green and orange. Different trajectories are shown which start at different initial conditions (b), (… view at source ↗
Figure 12
Figure 12. Figure 12: Passive and active optical bistability implementations. (a) Scheme of a Dynamically Reconfigurable Unified Microring Resonator (DRUM). Labels refer to phase shifter left/right (PSL/PSR), MZI left/right (MZL/MZR), and cavity energy field propa￾gating clockwise/counterclockwise (α1/α2). (b) Low-power linear transmission spectra for a phase imprinted by PSL of 0, π, and 2π as a function of the wavelength det… view at source ↗
Figure 13
Figure 13. Figure 13: Implementation of non-volatile memories. (a) Micro-cylinders GST pattern on a Si3N4 waveguide atop an SOI substrate. It is reprinted from [193] CC BY 4.0. (b) the layout of a photonic phase shifter based on ferroelectric BaTiO3 [194]. (c) Cross￾section of a memristor based on a MRR with an illustration of the device’s working mechanism, adapted from [195] CC BY 4.0. (d) Cross-section of the charge-trappin… view at source ↗
Figure 14
Figure 14. Figure 14: Overview of neural network architectures relevant to temporal memory processing and neuromorphic photonic imple￾mentations. A feed-forward neural network maps an input vector x through one (or more) hidden layers h to produce an output y, where information propagates only in the forward direction. A recurrent neural network includes feedback connections, so the in￾ternal state h(t) depends on the current … view at source ↗
Figure 15
Figure 15. Figure 15: Implementations of spatial RC in PICs. (a) Integrated spatial RC circuit composed of 2 × 2 combiners connected by delay lines (waveguides curled up into spirals) providing memory for high-speed signal processing (image reprinted with permission from [231], licensed under CC BY-NC-ND 3.0). (b) Schematic of an all-optical trainable readout for a photonic spatial RC system, where balanced MZIs and heater-bas… view at source ↗
Figure 1
Figure 1. Figure 1: Citation Kosuke Takano, Chihiro Sugano, Masanobu Inubushi, Kazuyuki Yoshimura, Satoshi Sunada, Kazutaka Kanno, Atsushi Uchida, "Compact reservoir computing with a photonic integrated circuit," Opt. Express 26, 29424-29439 (2018); https://opg.optica.org/oe/abstract.cfm?URI=oe-26-22-29424 Image © 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement and may be used for n… view at source ↗
Figure 16
Figure 16. Figure 16: Photonic implementations of delay-based reservoir computing. (a), (b) Integrated delay-based RC systems based on a distributed-feedback (DFB) laser, whose external cavities host optical delay lines (memory source) and amplifiers (images reprinted with permission respectively from [273] and [274] © Optical Society of America). (c) Delay-based photonic reservoir combining the nonlinear memory of an integrat… view at source ↗
Figure 17
Figure 17. Figure 17: Hybrid spatial and delay-based reservoir computing architectures. (a) A passive PIC hosts a combination of spatial and delay-based RC. The input is encoded as optical pulses, masking is performed on-chip exploiting a programmable feed￾forward network with multiple delay lines and a reservoir is implemented through coupled programmable optical cavities formed by waveguide loops (image from [266], licensed … view at source ↗
Figure 18
Figure 18. Figure 18: Response-induced nonlinear reservoir computing implementations in PICs. (a) Schematic representation of a simulated photonic reservoir based on 9 nonlinear coupled MRRs employed in [291]. (b) A single nonlinear silicon MRR is exploited for delay-based RC. A pump and probe mechanism is employed, via two input optical wavelengths that are nonlinearly coupled by the MRR’s nonlinear dynamics (image from [167]… view at source ↗
Figure 19
Figure 19. Figure 19: Implementation of PNNs with recurrent dynamics and adaptive memory. (a) A photonic RNN where on optic-electric￾optic (O-E-O) conversion is employed for non-coherent summation and nonlinearity in the artificial neurons. A single feedback waveguide takes the role of multiple recurrent connections by exploiting WDM. Synaptic weights are applied to specific wavelength channels via tunable MRRs (image from [28… view at source ↗
Figure 20
Figure 20. Figure 20: Implementations of a photonic Reservoir Computing for optical signal equalization. (a) Semiconductor laser with time-delayed optical feedback [301]. The simulated signal after fiber-optic transmission is randomly masked offline and applied to the injection laser. Then, this is directed to the response laser, which receives delayed optical feedback from a fiber loop providing a delay τ = 66 ns. Optical swi… view at source ↗
Figure 21
Figure 21. Figure 21: Time-delayed complex perceptron. (a) Schematics of an N-channel Time-Delayed Complex Perceptron [289, 308, 268, 290, 309]. The optical input time sequence x(t) is split into N copies, which then propagate through spiralized waveguides to accumulate a relative delay multiple of ∆t. Copies traveling in each channel are then applied with independent complex weights wi (i = 1, . . . , N). Finally, they reach … view at source ↗
Figure 22
Figure 22. Figure 22: Schematics of a tunable lattice filter implemented in SiN and used in [310, 311] for CD equalization. The input signal propagates through subunits constituted by a tunable coupler and an unbalanced MZI. The imbalance is provided by delay lines implemented via spirals of lengths ∆L (first and last units) or 2∆L. The inset shows the internal structure of a tunable coupler, which is represented by a balanced… view at source ↗
Figure 23
Figure 23. Figure 23: Schematics of a cascade of N MRRs used in [312, 313] for CD equalization. As shown in the inset, within each subunit, the coupling between the bus waveguide and the resonant structure is provided by a MZI, which can be thermally tuned to regulate the corresponding coupling coefficient κi with i = 1, . . . , N. Each MRR is then equipped with another phase shifter that drives the phase weight ϕi. Other filt… view at source ↗
read the original abstract

The rapid scaling of artificial neural networks has exposed fundamental limitations of conventional von Neumann computing architectures. In these systems, the physical separation between memory and processing creates a bottleneck, as computational capabilities outpace the ability of memory and interconnects to supply and retrieve data. In contrast, biological neural systems inherently co-localize computation and memory through distributed, dynamical processes. Neuromorphic computing seeks to emulate this paradigm by leveraging physical substrates whose intrinsic dynamics simultaneously encode and process information. Among emerging platforms, silicon photoncis offer a compelling approach due to its high bandwidth, low-loss propagation, and inherent parallelism. This review examines the role of memory in integrated photonic neuromorphic systems, with emphasis on the physical mechanisms that provide volatile (short-term) and non-volatile (long-term) memory in silicon-on-insulator and hybrid silicon-on-insulator platforms. Drawing inspiration from digital, biological, and photonic memory architectures, we classify existing approaches based on their underlying physical principles. We cover implementations ranging from delay lines and slow-light structures to multistable dynamics and structural memory based on charge trapping and phase-change materials. We then discuss how these mechanisms support photonic neural network architectures, including feed-forward, reservoir computing, spiking and hybrid optoelectronic recurrent systems, and assess their relevance for time-dependent singal-processing tasks such as channel equalization in telecommunications. This review aims to establish a unified framework for understanding memory and learning in neuromorphic photonics and outlines key challenges and opportunities for scalable, energy-efficient neuromorphic hardware.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The paper is a review that classifies memory mechanisms in integrated photonic neuromorphic systems according to their underlying physical principles, covering volatile and non-volatile approaches such as delay lines, slow-light structures, multistable dynamics, charge trapping, and phase-change materials in silicon-on-insulator and hybrid platforms. It connects these mechanisms to architectures including feed-forward networks, reservoir computing, spiking systems, and hybrid optoelectronic recurrent networks, with discussion of applications like time-dependent signal processing (e.g., channel equalization), and concludes by outlining a unified framework along with challenges and opportunities for scalable, energy-efficient hardware.

Significance. If the classification and framework hold, the review is significant as a structured synthesis that bridges physical mechanisms with neuromorphic photonic architectures, providing a reference point for researchers working on co-localized memory and computation in optics. It explicitly credits the diversity of existing implementations across platforms and highlights practical relevance for telecommunications tasks without claiming new experimental results.

minor comments (3)
  1. §3 (classification of mechanisms): the distinction between 'structural memory' via charge trapping and phase-change materials could be clarified with a brief table comparing volatility timescales and integration compatibility, as the current prose description risks conflating short-term and long-term behaviors in hybrid platforms.
  2. §4.2 (reservoir computing architectures): the discussion of delay-line-based reservoirs would benefit from explicit citation of at least two recent experimental benchmarks on energy per operation to strengthen the claim of 'low-loss propagation' advantages over electronic counterparts.
  3. Abstract and §5 (challenges): the phrase 'prohibitive losses or complexity' in the context of scalability is used without a quantitative threshold; adding a short paragraph with example loss budgets from cited SOI works would improve precision.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of our review, accurate summary of its scope, and recommendation for minor revision. We are pleased that the classification of memory mechanisms and the proposed unified framework are viewed as a useful synthesis bridging physical principles with neuromorphic photonic architectures.

Circularity Check

0 steps flagged

No significant circularity: review paper with no derivations

full rationale

This is a survey/review paper whose stated purpose is to classify existing memory mechanisms (delay lines, slow-light, multistable dynamics, charge trapping, phase-change materials) in silicon-on-insulator and hybrid platforms and to outline challenges for neuromorphic photonic architectures. No original equations, predictions, fitted parameters, or derivation chains are presented in the abstract or described scope. The central claim reduces to literature synthesis and classification rather than any self-referential construction, self-citation load-bearing argument, or renaming of results. The paper is therefore self-contained as a review and exhibits no circularity by the defined criteria.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

This is a review article summarizing existing research. No new free parameters, axioms, or invented entities are introduced by the authors.

pith-pipeline@v0.9.0 · 5591 in / 1006 out tokens · 60031 ms · 2026-05-08T10:29:24.849438+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

300 extracted references · 10 canonical work pages · 1 internal anchor

  1. [1]

    Deep learning

    Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. “Deep learning”. In:nature521.7553 (2015), pp. 436– 444

    2015

  2. [2]

    A comprehensive review of deep learning: Architectures, recent advances, and applications

    Ibomoiye Domor Mienye and Theo G Swart. “A comprehensive review of deep learning: Architectures, recent advances, and applications”. In:Information15.12 (2024), p. 755

    2024

  3. [3]

    Gordon E Moore et al.Cramming more components onto integrated circuits. 1965

    1965

  4. [4]

    Design of ion-implanted MOSFET’s with very small physical dimensions

    Robert H Dennard et al. “Design of ion-implanted MOSFET’s with very small physical dimensions”. In:IEEE Journal of solid-state circuits9.5 (2003), pp. 256–268

    2003

  5. [5]

    Dark silicon and the end of multicore scaling

    Hadi Esmaeilzadeh et al. “Dark silicon and the end of multicore scaling”. In:Proceedings of the 38th annual international symposium on Computer architecture. 2011, pp. 365–376

    2011

  6. [6]

    1.1 computing’s energy problem (and what we can do about it)

    Mark Horowitz. “1.1 computing’s energy problem (and what we can do about it)”. In:2014 IEEE international solid-state circuits conference digest of technical papers (ISSCC). IEEE. 2014, pp. 10–14

    2014

  7. [7]

    Attention is all you need

    Ashish Vaswani et al. “Attention is all you need”. In:Advances in neural information processing systems30 (2017)

    2017

  8. [8]

    Ai and memory wall

    Amir Gholami et al. “Ai and memory wall”. In:IEEE Micro44.3 (2024), pp. 33–39

    2024

  9. [9]

    Eric R Kandel et al.Principles of neural science. V ol. 4. McGraw-hill New York, 2000

    2000

  10. [10]

    Memory engram cells have come of age

    Susumu Tonegawa et al. “Memory engram cells have come of age”. In:Neuron87.5 (2015), pp. 918–931

    2015

  11. [11]

    Neuromorphic electronic systems

    Carver Mead. “Neuromorphic electronic systems”. In:Proceedings of the IEEE78.10 (2002), pp. 1629–1636

    2002

  12. [12]

    Memory and information processing in neuromorphic systems

    Giacomo Indiveri and Shih-Chii Liu. “Memory and information processing in neuromorphic systems”. In: Proceedings of the IEEE103.8 (2015), pp. 1379–1397

    2015

  13. [13]

    Neuromorphic computing at scale

    Dhireesha Kudithipudi et al. “Neuromorphic computing at scale”. In:Nature637.8047 (2025), pp. 801–812

    2025

  14. [14]

    Femtojoule per MAC Neuromorphic Photonics: An Energy and Technology Roadmap,

    Daniel Brunner et al. “Roadmap on neuromorphic photonics”. In:arXiv preprint arXiv:2501.07917(2025)

    work page arXiv 2025

  15. [15]

    Computer science as empirical inquiry: Symbols and search

    Allen Newel and Herbert A Simon. “Computer science as empirical inquiry: Symbols and search”. In:Com- munications of the ACM19.3 (1976), pp. 113–126

    1976

  16. [16]

    A modern approach

    Stuart Russell, Peter Norvig, and Artificial Intelligence. “A modern approach”. In:Artificial Intelligence. Prentice-Hall, Egnlewood Cliffs25.27 (1995), pp. 79–80

    1995

  17. [17]

    Neural networks and physical systems with emergent collective computational abilities

    John J Hopfield. “Neural networks and physical systems with emergent collective computational abilities.” In: Proceedings of the national academy of sciences79.8 (1982), pp. 2554–2558

    1982

  18. [18]

    Learning internal representations by error propagation

    David E. Rumelhart, Geoffrey E. Hinton, and Ronald J. Williams. “Learning internal representations by error propagation”. In: 1986.URL:https://api.semanticscholar.org/CorpusID:62245742

    1986

  19. [19]

    Photonics for artificial intelligence and neuromorphic computing

    Bhavin J Shastri et al. “Photonics for artificial intelligence and neuromorphic computing”. In:Nature Photon- ics15.2 (2021), pp. 102–114

    2021

  20. [20]

    Experimental realization of any discrete unitary operator

    Michael Reck et al. “Experimental realization of any discrete unitary operator”. In:Physical review letters 73.1 (1994), p. 58

    1994

  21. [21]

    Deep learning with coherent nanophotonic circuits

    Yichen Shen et al. “Deep learning with coherent nanophotonic circuits”. In:Nature photonics11.7 (2017), pp. 441–446

    2017

  22. [22]

    John Wiley & Sons, 2012

    Govind P Agrawal.Fiber-optic communication systems. John Wiley & Sons, 2012

    2012

  23. [23]

    All-optical computing with beyond 100-GHz clock rates

    Gordon HY Li et al. “All-optical computing with beyond 100-GHz clock rates”. In:arXiv preprint arXiv:2501.05756(2025)

    work page arXiv 2025

  24. [24]

    Optical nonlinearity, bistability, and signal processing in semi- conductors

    Nasser Peyghambarian and Hyatt M Gibbs. “Optical nonlinearity, bistability, and signal processing in semi- conductors”. In:Journal of the Optical Society of America B2.7 (1985), pp. 1215–1227

    1985

  25. [25]

    Holography in artificial neural networks

    Demetri Psaltis et al. “Holography in artificial neural networks”. In:Nature343.6256 (1990), pp. 325–330

    1990

  26. [26]

    Analog optical computing

    Daniel R Solli and Bahram Jalali. “Analog optical computing”. In:Nature Photonics9.11 (2015), pp. 704– 706

    2015

  27. [27]

    Taylor & Francis, 2016

    Laurent Vivien and Lorenzo Pavesi.Handbook of silicon photonics. Taylor & Francis, 2016. 54 Memory in Integrated Photonic Neural Networks: From Physical Mechanisms to Neuromorphic Architectures

    2016

  28. [28]

    CRC press, 2017

    Paul R Prucnal and Bhavin J Shastri.Neuromorphic photonics. CRC press, 2017

    2017

  29. [29]

    An introduction to InP-based generic integration technology

    Meint Smit et al. “An introduction to InP-based generic integration technology”. In:Semiconductor Science and Technology29.8 (2014), p. 083001

    2014

  30. [30]

    Integrated photonics on thin-film lithium niobate

    Di Zhu et al. “Integrated photonics on thin-film lithium niobate”. In:Advances in Optics and Photonics13.2 (2021), pp. 242–352

    2021

  31. [31]

    The past, present, and future of silicon photonics

    Richard Soref. “The past, present, and future of silicon photonics”. In:IEEE Journal of selected topics in quantum electronics12.6 (2006), pp. 1678–1687

    2006

  32. [32]

    Silicon optical modulators

    Graham T Reed et al. “Silicon optical modulators”. In:Nature photonics4.8 (2010), pp. 518–526

    2010

  33. [33]

    Roadmapping the next generation of silicon photonics

    Sudip Shekhar et al. “Roadmapping the next generation of silicon photonics”. In:Nature Communications 15.1 (2024), p. 751

    2024

  34. [34]

    Hybrid and heterogeneous photonic integration

    Paramjeet Kaur et al. “Hybrid and heterogeneous photonic integration”. In:APL photonics6.6 (2021)

    2021

  35. [35]

    Psychology press, 2005

    Donald Olding Hebb.The organization of behavior: A neuropsychological theory. Psychology press, 2005

    2005

  36. [36]

    Dynamic analysis and reservoir computing applica- tion of a nonlinear microring resonator

    Stefano Gretter, Mattia Mancinelli, and Lorenzo Pavesi. “Dynamic analysis and reservoir computing applica- tion of a nonlinear microring resonator”. In:ACS photonics12.9 (2025), pp. 4956–4966

    2025

  37. [37]

    All-optical spiking neurosynaptic networks with self-learning capabilities

    Johannes Feldmann et al. “All-optical spiking neurosynaptic networks with self-learning capabilities”. In: Nature569.7755 (2019), pp. 208–214

    2019

  38. [38]

    Computation at the edge of chaos: Phase transitions and emergent computation

    Chris G Langton. “Computation at the edge of chaos: Phase transitions and emergent computation”. In:Phys- ica D: nonlinear phenomena42.1-3 (1990), pp. 12–37

    1990

  39. [39]

    Stable Hebbian learning from spike timing- dependent plasticity

    Mark CW Van Rossum, Guo Qiang Bi, and Gina G Turrigiano. “Stable Hebbian learning from spike timing- dependent plasticity”. In:Journal of neuroscience20.23 (2000), pp. 8812–8821

    2000

  40. [40]

    Structured chaos shapes spike-response noise entropy in balanced neural networks

    Guillaume Lajoie, Jean-Philippe Thivierge, and Eric Shea-Brown. “Structured chaos shapes spike-response noise entropy in balanced neural networks”. In:Frontiers in computational neuroscience8 (2014), p. 123

    2014

  41. [41]

    Integrated all-photonic non-volatile multi-level memory

    Carlos R ´ıos et al. “Integrated all-photonic non-volatile multi-level memory”. In:Nature photonics9.11 (2015), pp. 725–732

    2015

  42. [42]

    Exploring the potential of self-pulsing optical microresonators for spiking neural networks and sensing

    Stefano Biasi et al. “Exploring the potential of self-pulsing optical microresonators for spiking neural networks and sensing”. In:Communications Physics7.1 (2024), p. 380

    2024

  43. [43]

    Emergent Self-Adaptation in an Integrated Photonic Neural Network for Backpropagation-Free Learning

    Alessio Lugnan et al. “Emergent Self-Adaptation in an Integrated Photonic Neural Network for Backpropagation-Free Learning”. In:Advanced Science12.2 (2025), p. 2404920

    2025

  44. [44]

    Progress in Optoelectronic Synapses for Reservoir Computing: Materials, Device Inte- gration, and Neuromorphic System Applications

    Yongsheng Lei et al. “Progress in Optoelectronic Synapses for Reservoir Computing: Materials, Device Inte- gration, and Neuromorphic System Applications”. In:Laser & Photonics Reviews(2025), e02338

    2025

  45. [45]

    Integrated photonic synapses, neurons, memristors, and neural networks for photonic neu- romorphic computing

    Shufei Han et al. “Integrated photonic synapses, neurons, memristors, and neural networks for photonic neu- romorphic computing”. In:Opto-Electronic Technology1.3 (2025), pp. 1–17

    2025

  46. [46]

    Intelligent nanophotonics: when machine learning sheds light

    Nanfan Wu et al. “Intelligent nanophotonics: when machine learning sheds light”. In:eLight5.1 (2025), p. 5

    2025

  47. [47]

    Algorithms, architectures, and platform implementations of inte- grated photonic neural networks

    Xinyi Wang, Kun Liao, and Xiaoyong Hu. “Algorithms, architectures, and platform implementations of inte- grated photonic neural networks”. In:Applied Physics Reviews13.1 (2026)

    2026

  48. [48]

    Elements of episodic memory

    Endel Tulving. “Elements of episodic memory”. In: (1983)

    1983

  49. [49]

    Larry R Squire and Eric R Kandel.Memory: From mind to molecules. V ol. 69. Macmillan, 2003

    2003

  50. [50]

    Charles R Gallistel.The organization of learning.The MIT Press, 1990

    1990

  51. [51]

    Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path

    Tim VP Bliss and Terje Lømo. “Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path”. In:The Journal of physiology232.2 (1973), pp. 331–356

    1973

  52. [52]

    The molecular biology of memory storage: a dialogue between genes and synapses

    Eric R Kandel. “The molecular biology of memory storage: a dialogue between genes and synapses”. In: Science294.5544 (2001), pp. 1030–1038

    2001

  53. [53]

    Synaptic computation

    Larry F Abbott and Wade G Regehr. “Synaptic computation”. In:Nature431.7010 (2004), pp. 796–803

    2004

  54. [54]

    Working memory: Theories, models, and controversies

    Alan Baddeley. “Working memory: Theories, models, and controversies”. In:Annual review of psychology 63.1 (2012), pp. 1–29

    2012

  55. [55]

    Neuroscience-inspired artificial intelligence

    Demis Hassabis et al. “Neuroscience-inspired artificial intelligence”. In:Neuron95.2 (2017), pp. 245–258

    2017

  56. [56]

    Toward an integration of deep learning and neuroscience

    Adam H Marblestone, Greg Wayne, and Konrad P Kording. “Toward an integration of deep learning and neuroscience”. In:Frontiers in computational neuroscience10 (2016), p. 215943

    2016

  57. [57]

    Towards biologically plausible deep learning

    Yoshua Bengio et al. “Towards biologically plausible deep learning”. In:arXiv preprint arXiv:1502.04156 (2015)

    work page arXiv 2015

  58. [58]

    Backpropagation and the brain

    Timothy P Lillicrap et al. “Backpropagation and the brain”. In:Nature Reviews Neuroscience21.6 (2020), pp. 335–346. 55 Memory in Integrated Photonic Neural Networks: From Physical Mechanisms to Neuromorphic Architectures

    2020

  59. [59]

    Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type

    Guo-qiang Bi and Mu-ming Poo. “Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type”. In:Journal of neuroscience18.24 (1998), pp. 10464–10472

    1998

  60. [60]

    The neurobiology of consolidations, or, how stable is the engram?

    Yadin Dudai. “The neurobiology of consolidations, or, how stable is the engram?” In:Annu. Rev. Psychol.55.1 (2004), pp. 51–86

    2004

  61. [61]

    Synapses and memory storage

    Mark Mayford, Steven A Siegelbaum, and Eric R Kandel. “Synapses and memory storage”. In:Cold Spring Harbor perspectives in biology4.6 (2012), a005751

    2012

  62. [62]

    Memory formation depends on both synapse-specific modifications of synaptic strength and cell-specific increases in excitability

    John Lisman et al. “Memory formation depends on both synapse-specific modifications of synaptic strength and cell-specific increases in excitability”. In:Nature neuroscience21.3 (2018), pp. 309–314

    2018

  63. [63]

    The molecular and systems biology of memory

    Eric R Kandel, Yadin Dudai, and Mark R Mayford. “The molecular and systems biology of memory”. In:Cell 157.1 (2014), pp. 163–186

    2014

  64. [64]

    Memory consolidation

    Larry R Squire et al. “Memory consolidation”. In:Cold Spring Harbor perspectives in biology7.8 (2015), a021766

    2015

  65. [65]

    Human memory: A proposed system and its control processes

    Richard C Atkinson and Richard M Shiffrin. “Human memory: A proposed system and its control processes”. In:Psychology of learning and motivation. V ol. 2. Elsevier, 1968, pp. 89–195

    1968

  66. [66]

    Working memory: The interface between memory and cognition

    Alan Baddeley. “Working memory: The interface between memory and cognition”. In:Journal of cognitive neuroscience4.3 (1992), pp. 281–288

    1992

  67. [67]

    Neuron activity related to short-term memory

    Joaquin M Fuster and Garrett E Alexander. “Neuron activity related to short-term memory”. In:Science 173.3997 (1971), pp. 652–654

    1971

  68. [68]

    Synaptic mechanisms and network dynamics underlying spatial working memory in a cortical network model

    Albert Compte et al. “Synaptic mechanisms and network dynamics underlying spatial working memory in a cortical network model”. In:Cerebral cortex10.9 (2000), pp. 910–923

    2000

  69. [69]

    Working memory: Past, present. . . and future

    AD Baddeley and GJ Hitch. “Working memory: Past, present. . . and future”. In:The cognitive neuroscience of working memory(2007), pp. 1–20

    2007

  70. [70]

    Cellular basis of working memory

    Patricia S Goldman-Rakic. “Cellular basis of working memory”. In:Neuron14.3 (1995), pp. 477–485

    1995

  71. [71]

    An integrative theory of prefrontal cortex function

    Earl K Miller and Jonathan D Cohen. “An integrative theory of prefrontal cortex function”. In:Annual review of neuroscience24.1 (2001), pp. 167–202

    2001

  72. [72]

    Working memory in the prefrontal cortex

    Shintaro Funahashi. “Working memory in the prefrontal cortex”. In:Brain sciences7.5 (2017), p. 49

    2017

  73. [73]

    Structural changes accompanying memory storage

    Craig H Bailey and Eric R Kandel. “Structural changes accompanying memory storage.” In:Annual review of physiology(1993)

    1993

  74. [74]

    The organization of recent and remote memories

    Paul W Frankland and Bruno Bontempi. “The organization of recent and remote memories”. In:Nature reviews neuroscience6.2 (2005), pp. 119–130

    2005

  75. [75]

    Memory systems of the brain: a brief history and current perspective

    Larry R Squire. “Memory systems of the brain: a brief history and current perspective”. In:Neurobiology of learning and memory82.3 (2004), pp. 171–177

    2004

  76. [76]

    A neostriatal habit learning system in humans

    Barbara J Knowlton, Jennifer A Mangels, and Larry R Squire. “A neostriatal habit learning system in humans”. In:Science273.5280 (1996), pp. 1399–1402

    1996

  77. [77]

    Actions and habits: the development of behavioural autonomy

    Anthony Dickinson. “Actions and habits: the development of behavioural autonomy”. In:Philosophical Trans- actions of the Royal Society of London. B, Biological Sciences308.1135 (1985), pp. 67–78

    1985

  78. [78]

    Episodic and semantic memory

    Endel Tulving et al. “Episodic and semantic memory”. In:Organization of memory1.381-403 (1972), p. 1

    1972

  79. [79]

    Imaging cognition II: An empirical review of 275 PET and fMRI studies

    Roberto Cabeza and Lars Nyberg. “Imaging cognition II: An empirical review of 275 PET and fMRI studies”. In:Journal of cognitive neuroscience12.1 (2000), pp. 1–47

    2000

  80. [80]

    Transforming experiences: Neurobiology of memory updating/editing

    Daniel Osorio-G ´omez et al. “Transforming experiences: Neurobiology of memory updating/editing”. In:Fron- tiers in Systems Neuroscience17 (2023), p. 1103770

    2023

Showing first 80 references.