Real-Time Computing Without Stable States: A New Framework for Neural Computation Based on Perturbations
Open accessThe DOI matches its Crossref record; 28 of the 28 references it cites were found in a registry or an index.
Abstract
A key challenge for neural modeling is to explain how a continuous stream of multimodal input from a rapidly changing environment can be processed by stereotypical recurrent circuits of integrate-and-fire neurons in real time. We propose a new computational model for real-time computing on time-varying input that provides an alternative to paradigms based on Turing machines or attractor neural networks. It does not require a task-dependent construction of neural circuits. Instead, it is based on principles of high-dimensional dynamical systems in combination with statistical learning theory and can be implemented on generic evolved or found recurrent circuitry. It is shown that the inherent transient dynamics of the high-dimensional dynamical system formed by a sufficiently large and heterogeneous neural circuit may serve as universal analog fading memory. Readout neurons can learn to extract in real time from the current state of such recurrent neural circuit information about current and past inputs that may be needed for diverse tasks. Stable internal states are not required for giving a stable output, since transient internal states can be transformed by readout neurons into stable target outputs due to the high dimensionality of the dynamical system. Our approach is based on a rigorous computational model, the liquid state machine, that, unlike Turing machines, does not require sequential transitions between well-defined discrete internal states. It is supported, as the Turing machine is, by rigorous mathematical results that predict universal computational power under idealized conditions, but for the biologically more realistic scenario of real-time processing of time-varying inputs. Our approach provides new perspectives for the interpretation of neural coding, the design of experiments and data analysis in neurophysiology, and the solution of problems in robotics and neurotechnology.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Although work in artificial intelligence and machine-learning research has shown that similar or even more complex outcomes can be achieved with a variety of techniques, including deep neural networks with trained weights (Gibney, 2015), reservoir networks with a trained output mapping (Maass et al., 2002;Jaeger & Haas, 2004), networks that are evolved using a genetic algorithm (Iizuka & Di Paolo,”
From A potential mechanism for Gibsonian resonance: behavioral entrainment emerges from local homeostasis in an unsupervised reservoir network (2024) · cites this paper as Maass et al., 2002
In the citing paper
“Finally, we show that a randomly connected circuit model, akin to a “reservoir” network13-15, can reproduce the observed distribution of timescales, provided that the network operates at the critical point (or “edge of chaos”)16-18.”
From A reservoir of time constants for memory traces in cortical neurons (2011) · cites this paper as [13]
In the citing paper
“Reservoir-type networks have been the subject of active research in computational neuroscience and machine learning13-15, but so far experimental support that such networks are adopted by the brain has been lacking.”
From A reservoir of time constants for memory traces in cortical neurons (2011) · cites this paper as [13]
In the citing paper
“Those models predict that the memory of input signals is stored in a large, recurrent and heterogeneous network (reservoir), in a distributed manner, and that a desired output is obtained by a trainable combination of the response signals in the reservoir.”
From A reservoir of time constants for memory traces in cortical neurons (2011) · cites this paper as [13]
Citations
Cites
Wu, Y. (1999) · Technometrics · doi:10.1080/00401706.1999.10485951 · Cited by 27,221
Introduction to the theory of neural computation
(1994) · Neural Networks · doi:10.1016/0893-6080(94)90014-0 · Cited by 6,401
The Synaptic Organization of the Brain
(2004) · Oxford University Press eBooks · doi:10.1093/acprof:oso/9780195159561.001.1 · Cited by 3,546 · cited 2×
“This informationisprocessed by extremely complexbut surprisingly stereotypic microcircuits that can perform a wide spectrum of tasks (Shep-herd, 1988 ; Douglas & Martin, 1998 ; von Melchner, Pallas, & Sur, 2000).”
Dynamics of ongoing activity: explanation of the large variability in evoked cortical responses
Arieli, A., Sterkin, A., Grinvald, A., & Aertsen, A. (1996) · Science · doi:10.1126/science.273.5283.1868 · Cited by 1,660 · cited 1×
“Furthermore, although the input rate may contain salient information that is constant for a particular readout element, it may not be for another (see, for example, Figure 8), indicating that equivalence classes and dynamic stability exist purely from the perspective of the readout elements.”
BackgroundDifferential signaling via the same axon of neocortical pyramidal neurons
Markram, H., Wang, Y., & Tsodyks, M. (1998) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.95.9.5323 · Cited by 1,102 · cited 2×
“In the case of a synaptic connection from a to b , we modeled the synaptic dynamics according to the model proposed in Markram, Wang, and Tsodyks (1998), with the synaptic parameters U (use), D (time constant for depression), and F (time constant for facilitation) randomly chosen from gaussian…”
BackgroundMethodSelf-citationOrganizing principles for a diversity of GABAergic interneurons and synapses in the neocortex
Gupta, A., Wang, Y., & Markram, H. (2000) · Science · doi:10.1126/science.287.5451.273 · Cited by 1,036
Self-citationFading memory and the problem of approximating nonlinear operators with Volterra series
Boyd, S., & Chua, L. (1985) · IEEE Transactions on Circuits and Systems · doi:10.1109/tcs.1985.1085649 · Cited by 991 · cited 2×
“> The proof of this theorem follows from the Stone-Weierstrass approximation theorem, like the proof of theorem 1 in Boyd and Chua (1985).”
BackgroundGradient calculations for dynamic recurrent neural networks: a survey
Pearlmutter, B. A. (1995) · IEEE Transactions on Neural Networks · doi:10.1109/72.410363 · Cited by 582 · cited 1×
“Methods for controllingthe dynamics of recurrent neural networks through adaptive mechanisms are reviewed in Pearlmutter (1995).”
Temporal Information Transformed into a Spatial Code by a Neural Network with Realistic Properties
Buonomano, D. V., & Merzenich, M. M. (1995) · Science · doi:10.1126/science.7863330 · Cited by 449 · cited 1×
“Special cases of this mechanism have already been reported in Buonomano and Merzenich (1995) and Dominey, Arbib, and Joseph (1995).”
BackgroundLearning and Extracting Finite State Automata with Second-Order Recurrent Neural Networks
Giles, C. L., Miller, C. B., Chen, D., Chen, H. H., Sun, G. Z., & Lee, Y. C. (1992) · Neural Computation · doi:10.1162/neco.1992.4.3.393 · Cited by 392
Cited by
Items 601–700 of 1,014
Fluid network dynamics in the prefrontal cortex during multiple strategy switching
Malagon‐Vina, H., Ciocchi, S., Passecker, J., Dorffner, G., & Klausberger, T. (2018) · Nature Communications · doi:10.1038/s41467-017-02764-x · Cited by 63
Delayed dynamical systems: networks, chimeras and reservoir computing
Hart, J. D., Larger, L., Murphy, T. E., & Roy, R. (2019) · Philosophical Transactions of the Royal Society A Mathematical Physical and Engineering Sciences · doi:10.1098/rsta.2018.0123 · Cited by 68
“They were ‘temporally’ discovered and invented, respectively, in the early 2000s [38,66,67], and they were ‘geographically’ connected to delay dynamics during the Delayed Complex Systems conference DCS’12, a decade later.”
BackgroundMethodGottwald, G. A., & Reich, S. (2021) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0066080 · Cited by 71
“The framework of random feature maps was extended to include internal dynamics in so called echo-state networks and reservoir computers with remarkable success in forecasting dynamical systems (Maass et al., 2002; Jaeger, 2002; Jaeger and Haas, 2004; Pathak et al., 2018; Jüngling et al., 2019; Algar et al., 2019; Nadi…”
Spike-Timing Dependent Plasticity for Evolved Robots
Paolo, E. A. D. (2002) · Adaptive Behavior · doi:10.1177/1059712302010003006 · Cited by 59
Radar-Based Hand Gesture Recognition Using Spiking Neural Networks
Tsang, I. J., Corradi, F., Sifalakis, M., Leekwijck, W. V., & Latré, S. (2021) · Electronics · doi:10.3390/electronics10121405 · Cited by 67
“The network will be dominated by the external stimuli, degrading the contribution from the recurrent network and impacting the fading memory capability of the LSM, which is a fundamental property for the universal computational power of the LSM [18,19].”
BackgroundMethodInfluentialFerroelectric materials for neuroinspired computing applications
Wang, D., Hao, S., Dkhil, B., Tian, B., & Duan, C. (2023) · Fundamental Research · doi:10.1016/j.fmre.2023.04.013 · Cited by 63
“Apart from LSTM-based RNNs, researchers have proposed a simpli-fied RNN form called reservoir computing (RC), which was developed from the echo state network [95] and liquid state machine [96] .”
Predicting turbulent dynamics with the convolutional autoencoder echo state network
Racca, A., Doan, N. A. K., & Magri, L. (2023) · Journal of Fluid Mechanics · doi:10.1017/jfm.2023.716 · Cited by 72
Improving liquid state machines through iterative refinement of the reservoir
Norton, D. P., & Ventura, D. A. (2010) · Neurocomputing · doi:10.1016/j.neucom.2010.08.005 · Cited by 67
“A metric for separation was first devised by Goodman [11] and is inspired by the description of the properties of a liquid presented by Maass [16].”
MethodOperating in a Reverberating Regime Enables Rapid Tuning of Network States to Task Requirements
Wilting, J., Dehning, J., Neto, J. P., Rudelt, L., Wibral, M., Zierenberg, J., & Priesemann, V. (2018) · Frontiers in Systems Neuroscience · doi:10.3389/fnsys.2018.00055 · Cited by 62
“This is at the heart of reservoir computing in echo state networks or liquid state machines [9, 10, 11, 12, 13, 5].”
BackgroundDeep Liquid State Machines With Neural Plasticity for Video Activity Recognition
Soures, N., & Kudithipudi, D. (2019) · Frontiers in Neuroscience · doi:10.3389/fnins.2019.00686 · Cited by 58
“The two algorithms are otherwise known as the Echo State Network (ESN) (Jaeger, 2001) and the Liquid State Machine (LSM) (Maass et al., 2002).”
MethodDeep-ESN: A Multiple Projection-encoding Hierarchical Reservoir Computing FrameworkPreprint
Ma, Q., Shen, L., & Cottrell, G. W. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1711.05255 · Cited by 58
“The two most well-known RC models are Liquid State Machines and Echo State Networks [2], [10], [11].”
MethodEdge of Chaos and Avalanches in Neural Networks with Heavy-Tailed Synaptic Weight Distribution
Kuśmierz, Ł., Ogawa, S., & Toyoizumi, T. (2020) · Physical Review Letters · doi:10.1103/physrevlett.125.028101 · Cited by 57
Chen, G., Scherr, F., & Maass, W. (2022) · Science Advances · doi:10.1126/sciadv.abq7592 · Cited by 65
Self-citationReservoir computing benchmarks: a tutorial review and critique
Wringe, C., Trefzer, M. A., & Stepney, S. (2025) · International Journal of Parallel Emergent and Distributed Systems · doi:10.1080/17445760.2025.2472211 · Cited by 77
Artificial Neural Network Models
Peter Tiňo, B., Beňušková, Ľ., & Sperduti, A. (2015) · Springer eBooks · doi:10.1007/978-3-662-43505-2_27 · Cited by 61
Reducing the phase sensitivity of laser-based optical reservoir computing systems
Nguimdo, R. M., Verschaffelt, G., Danckaert, J., & der Sande, G. V. (2016) · Optics Express · doi:10.1364/oe.24.001238 · Cited by 57
Matsubara, T., & Torikai, H. (2015) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2015.2425893 · Cited by 57
Neuromorphic Time-Dependent Pattern Classification with Organic Electrochemical Transistor Arrays
Pecqueur, S., Talamo, M. M., Guérin, D., Blanchard, P., Roncali, J., Vuillaume, D., & Alibart, F. (2018) · LillOA (Université de Lille (University Of Lille)) · Cited by 57
Röhm, A., Gauthier, D. J., & Fischer, I. (2021) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0065813 · Cited by 61
Coding and learning of behavioral sequences
Melamed, O., Gerstner, W., Maass, W., Tsodyks, M., & Markram, H. (2004) · Trends in Neurosciences · doi:10.1016/j.tins.2003.10.014 · Cited by 56
Self-citationSpatiotemporal Pattern Recognition via Liquid State Machines
Goodman, E. L., & Ventura, D. A. (2006) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2006.246880 · Cited by 62
“Here, we explore the application of liquid state machines (LSMs) [8] [9] [10] [11] [12] to spatiotemporal pattern recognition.”
Multi-reservoir echo state networks with sequence resampling for nonlinear time-series prediction
Li, Z., & Tanaka, G. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.08.122 · Cited by 65
Thermal Management in Neuromorphic Materials, Devices, and Networks
Torres, F., Basaran, A. C., & Schuller, I. K. (2022) · Advanced Materials · doi:10.1002/adma.202205098 · Cited by 61
The computational foundations of dynamic coding in working memory
Stroud, J. P., Duncan, J., & Lengyel, M. (2024) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2024.02.011 · Cited by 65
“Whether and how many attractor states exist in a network ’ s state space, or if instead a network exhibits attractor-free (e.g., chaotic [52,54,89]) dynamics, depends on the single neuron nonlinearity [ f ⋅ in Equation I] and, importantly, the setting of its synaptic connection strengths ( W ij in…”
Quantum cognition and bounded rationality
Blutner, R., & Graben, P. B. (2015) · Synthese · doi:10.1007/s11229-015-0928-5 · Cited by 60
“However, they can also be created internally, from an autoepistemic point of view, which has been shown in the framework of liquid computation (Maass et al. 2002; beim Graben et al. 2009).”
BackgroundRepresentation of Time-Varying Stimuli by a Network Exhibiting Oscillations on a Faster Time Scale
Shamir, M., Ghitza, O., Epstein, S., & Kopell, N. (2009) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1000370 · Cited by 58
“, [1–8]).”
Reservoir Computing as a Model for In-Materio Computing
Dale, M., Miller, J. F., & Stepney, S. (2016) · Emergence, complexity and computation · doi:10.1007/978-3-319-33924-5_22 · Cited by 61
“Network mediated separation is deemed fundamentally important for input time-series networks, with similarities to the separation property in LSMs (see [72]).”
BackgroundMethodInfluentialHart, A., Hook, J., & Dawes, J. (2021) · Physica D Nonlinear Phenomena · doi:10.1016/j.physd.2021.132882 · Cited by 64
Stable and dynamic representations of value in the prefrontal cortex
Enel, P., Wallis, J. D., & Rich, E. L. (2020) · eLife · doi:10.7554/elife.54313 · Cited by 59
“Another neural network framework, with random recurrent connections referred to as reservoir computing, subsequently demonstrated that short term memory is readily obtained from generic networks in which different inputs (stimuli) elicit distinct trajectories thereby encoding stimuli for a short duration with highly d…”
BackgroundYildiz, I. B., von Kriegstein, K., & Kiebel, S. J. (2013) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1003219 · Cited by 55
“Recently, so called reservoir computing techniques using recurrent neural networks have been used for speech recognition [18,51,89–92] and provide excellent recognition results.”
BackgroundUsing echo state networks for classification: A case study in Parkinson's disease diagnosis
Lacy, S. E., Smith, S. L., & Lones, M. A. (2018) · Artificial Intelligence in Medicine · doi:10.1016/j.artmed.2018.02.002 · Cited by 55
“Two common implementations of reservoir computing are Echo State Networks (ESNs) [2] and Liquid State Machines (LSMs) [6].”
BackgroundGalaris, E., Fabiani, G., Gallos, I., Kevrekidis, I. G., & Siettos, C. I. (2022) · Journal of Scientific Computing · doi:10.1007/s10915-022-01883-y · Cited by 68
Kira, S., Safaai, H., Morcos, A. S., Panzeri, S., & Harvey, C. D. (2023) · Nature Communications · doi:10.1038/s41467-023-37804-2 · Cited by 54
Optimal information loading into working memory explains dynamic coding in the prefrontal cortex
Stroud, J. P., Watanabe, K., Suzuki, T., Stokes, M. G., & Lengyel, M. (2023) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2307991120 · Cited by 66
Rössert, C., Dean, P., & Porrill, J. (2015) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1004515 · Cited by 61
“This however was almost exclusively done for generic networks where neurons are always connected by negative and positive weights [27] or include both mutual exciting and inhibiting neurons in the randomly connected reservoir networks [28].”
BackgroundSpatiotemporal features for asynchronous event-based data
Lagorce, X., Ieng, S.-H., Clady, X., Pfeiffer, M., & Benosman, R. B. (2015) · Frontiers in Neuroscience · doi:10.3389/fnins.2015.00046 · Cited by 55
“…the layer of ESNs that predict the visual input, but this restriction could be lifted by replacing ESNs with their spiking counter-parts, called Liquid State Machines (LSMs) (Maass et al., 2002), which are computationally at least equivalent to ESNs (Maass and Markram, 2004; Büsing et al., 2010).”
BackgroundA Systematic Method for Configuring VLSI Networks of Spiking Neurons
Neftci, E., Chicca, E., Indiveri, G., & Douglas, R. J. (2011) · Neural Computation · doi:10.1162/neco_a_00182 · Cited by 55
Nguimdo, R. M., Lacot, É., Jacquin, O., Hugon, O., der Sande, G. V., & de Chatellus, H. G. (2017) · Optics Letters · doi:10.1364/ol.42.000375 · Cited by 55
Learning nonlinear input–output maps with dissipative quantum systems
Chen, J., & Nurdin, H. I. (2019) · Quantum Information Processing · doi:10.1007/s11128-019-2311-9 · Cited by 78
Consistency in echo-state networks
Lymburn, T., Khor, A., Stemler, T., Corrêa, D. C., Small, M., & Jüngling, T. (2019) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/1.5079686 · Cited by 61
Dynamical Learning of Dynamics
Klos, C., Kossio, Y. F. K., Goedeke, S., Gilra, A., & Memmesheimer, R.-M. (2020) · Physical Review Letters · doi:10.1103/physrevlett.125.088103 · Cited by 61
Muscinelli, S. P., Gerstner, W., & Schwalger, T. (2019) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1007122 · Cited by 64
Ros, F. D., Ranzini, S. M., Bülow, H., & Zibar, D. (2020) · IEEE Journal of Selected Topics in Quantum Electronics · doi:10.1109/jstqe.2020.2975607 · Cited by 60
“The concepts of echo state network (ESN) [26] and liquid state machine (LSM) [27] were proposed independently as simpler alternatives to overcome these training challenges.”
MethodInteger Echo State Networks: Efficient Reservoir Computing for Digital Hardware
Kleyko, D. V., Frady, E. P., Kheffache, M., & Osipov, E. (2020) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.3043309 · Cited by 61
“This approach originally appeared in two similar architectures: liquid state machines [15] and ESNs [16], now referred to as RC [1].”
BackgroundReservoir Computing as a Tool for Climate Predictability Studies
Nadiga, B. (2021) · Journal of Advances in Modeling Earth Systems · doi:10.1029/2020ms002290 · Cited by 55
Adaptable Hamiltonian neural networks
Han, C.-D., Glaz, B. J., Haile, M., & Lai, Y. (2021) · Physical Review Research · doi:10.1103/physrevresearch.3.023156 · Cited by 55
Dendritic Computing: Branching Deeper into Machine Learning
Acharya, J., Basu, A., Legenstein, R. A., Limbacher, T., Poirazi, P., & Wu, X. (2021) · Neuroscience · doi:10.1016/j.neuroscience.2021.10.001 · Cited by 63
Analyzing the weight dynamics of recurrent learning algorithms
Schiller, U., & Steil, J. J. (2004) · Neurocomputing · doi:10.1016/j.neucom.2004.04.006 · Cited by 55
“This point of view is further supported by a formal treatment of the one-output case of APRL which reveals a functional division of the network which resembles more the ”echo state network” [4,5] and ”liquid state machine” approaches [6,7] than classical backpropagation.”
ResultEvolution of Associative Learning in Chemical Networks
McGregor, S., Vasas, V., Husbands, P., & Fernando, C. (2012) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1002739 · Cited by 56
“There are important parallels here to liquid state machines [26,27] and other reservoir machines [28] and to random projections [29] in machine learning: information capture is not necessarily the hardest part of information processing, and randomly constituted systems can often accumulate information in a usable fash…”
BackgroundPrinciples of Neuromorphic Photonics
Shastri, B. J., Tait, A. N., de Lima, T. F., Nahmias, M. A., Peng, H.-T., & Prucnal, P. R. (2018) · Springer eBooks · doi:10.1007/978-1-4939-6883-1_702 · Cited by 57
Nanophotonic Reservoir Computing With Photonic Crystal Cavities to Generate Periodic Patterns
Fiers, M., Vaerenbergh, T. V., wyffels, F., Verstraeten, D., Schrauwen, B., Dambre, J., & Bienstman, P. (2014) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2013.2274670 · Cited by 54
“Index Terms— Integrated optics, optical neural network, pattern generation, photonic reservoir computing, supervised learning.”
BackgroundA Generalized Rotate-and-Fire Digital Spiking Neuron Model and Its On-FPGA Learning
Matsubara, T., Torikai, H., & Hishiki, T. (2011) · IEEE Transactions on Circuits & Systems II Express Briefs · doi:10.1109/tcsii.2011.2161705 · Cited by 54
“, [3] and [21]; 3) development of a low-power and small-size implementation method of the network for a neural prosthesis technology; and 4) extension of the learning algorithm by using other parameter optimization methods such as genetic algorithm, tabu search, and particle swarm optimization.”
MethodBai, K., & Yi, Y. (2018) · ACM Journal on Emerging Technologies in Computing Systems · doi:10.1145/3264659 · Cited by 63
Breaking symmetries of the reservoir equations in echo state networks
Herteux, J., & Räth, C. (2020) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0028993 · Cited by 62
Generative Models of Brain Dynamics
Ramezanian-Panahi, M., Abrevaya, G., Gagnon-Audet, J.-C., Voleti, V., Rish, I., & Dumas, G. (2022) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2022.807406 · Cited by 64
“Reservoir Computing A reservoir computer (RC) (Maass et al., 2002) is an RNN with a reservoir of interconnected spiking neurons.”
BackgroundBelief Propagation in Networks of Spiking Neurons
Steimer, A., Maass, W., & Douglas, R. (2009) · Neural Computation · doi:10.1162/neco.2009.08-08-837 · Cited by 54
“rate without introducing dependencies on the distant past, which the liquid pool is unable to provide because of its fading memory (Maass et al., 2002).”
BackgroundMethodInfluentialSelf-citationLocal Interactions in Neural Networks Explain Global Effects in Gestalt Processing and Masking
Herzog, M. H., Ernst, U. A., Etzold, A., & Eurich, C. W. (2003) · Neural Computation · doi:10.1162/089976603322297304 · Cited by 54
Applications of Metaheuristics in Reservoir Computing Techniques: A Review
Bala, A., Ismail, I., Ibrahim, R., & Sait, S. M. (2018) · IEEE Access · doi:10.1109/access.2018.2873770 · Cited by 53
“They aim to interpret the primary computational characteristics of neural microcircuits [12].”
BackgroundTime-delayed reservoir computing based on an optically pumped spin VCSEL for high-speed processing
Yang, Y., Zhou, P., Mu, P., & Li, N. (2022) · Nonlinear Dynamics · doi:10.1007/s11071-021-07140-5 · Cited by 55
“In 2007, the concept of reservoir computing (RC) was proposed by Verstraeten et al. [5] as a joint name of the echo state network (ESN) [6] and liquid state machine (LSM) [7].”
BackgroundNamiki, W., Nishioka, D., Yamaguchi, Y., Tsuchiya, T., Higuchi, T., & Terabe, K. (2023) · Advanced Intelligent Systems · doi:10.1002/aisy.202300228 · Cited by 60
“(2) and the predicted output reconstructed from the reservoir computing system in the training phase, which is measured under a magnetic field of 169 mT at an interval of 5 ns.”
BackgroundInfluentialInferring unknown unknowns: Regularized bias-aware ensemble Kalman filter
Nóvoa, A., Racca, A., & Magri, L. (2023) · Computer Methods in Applied Mechanics and Engineering · doi:10.1016/j.cma.2023.116502 · Cited by 62
Roy, S., Banerjee, A., & Basu, A. (2014) · IEEE Transactions on Biomedical Circuits and Systems · doi:10.1109/tbcas.2014.2362969 · Cited by 57
“In LSM-DER, the liquid described in [9] is followed by the two neuronal cell architecture depicted in Fig.”
BackgroundMethodInfluentialEvaluation of memory capacity of spin torque oscillator for recurrent neural networks
Tsunegi, S., Taniguchi, T., Miwa, S., Nakajima, K., Yakushiji, K., Fukushima, A., Yuasa, S., & Kubota, H. (2018) · Japanese Journal of Applied Physics · doi:10.7567/jjap.57.120307 · Cited by 56
Designing spontaneous behavioral switching via chaotic itinerancy
Inoue, K., Nakajima, K., & Kuniyoshi, Y. (2020) · Science Advances · doi:10.1126/sciadv.abb3989 · Cited by 55
“In addition, this embedding process is accomplished by modifying fewer parameters using the method of reservoir computing (32, 33).”
MethodNuntalid, N., Dhoble, K., & Kasabov, N. K. (2011) · Lecture notes in computer science · doi:10.1007/978-3-642-24955-6_54 · Cited by 53
“We construct a reservoir having a small-world interconnectivity pattern as described in [9].”
MethodBuiatti, M., Papo, D., Baudonnière, P.-M., & van Vreeswijk, C. (2007) · Neuroscience · doi:10.1016/j.neuroscience.2007.02.048 · Cited by 58
Bayesian optimization of hyper-parameters in reservoir computingPreprint
Yperman, J., & Becker, T. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1611.05193 · Cited by 53
“Echo state networks were introduced independently by Herbert Jaeger [11] and Wolfgang Maass (under the name Liquid State Machines [12]).”
MethodNeuromorphic Computing for Interactive Robotics: A Systematic Review
Aitsam, M., Davies, S., & Nuovo, A. G. D. (2022) · IEEE Access · doi:10.1109/access.2022.3219440 · Cited by 52
“There are many examples of converting data directly from sensors [19], [20], [21], intelligent systems with controlling manipulators [11], [22], and robots [23], [24] [25].”
BackgroundComputational role of structure in neural activity and connectivity
Ostojic, S., & Fusi, S. (2024) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2024.03.003 · Cited by 60
Brain-Inspired Reservoir Computing Using Memristors with Tunable Dynamics and Short-Term Plasticity
Armendarez, N. X., Mohamed, A. S. I., Dhungel, A., Hossain, M. R., Hasan, M. S., & Najem, J. S. (2024) · ACS Applied Materials & Interfaces · doi:10.1021/acsami.3c16003 · Cited by 54
Mean-field theory of echo state networks
Massar, M., & Massar, S. (2013) · Physical Review E · doi:10.1103/physreve.87.042809 · Cited by 58
Memristor Models for Machine Learning
Carbajal, J. P., Dambre, J., Hermans, M., & Schrauwen, B. (2015) · Neural Computation · doi:10.1162/neco_a_00694 · Cited by 54
“Reservoir computing (RC) is a supervised learning framework rooted in recurrent neural network research (seminal work in [5; 6], recent developments in [7; 8]).”
BackgroundSajda, P., Philiastides, M. G., & Parra, L. C. (2009) · IEEE Reviews in Biomedical Engineering · doi:10.1109/rbme.2009.2034535 · Cited by 49
“The basic notion of what is known as reservoir computing [43], [44] is to represent a stimulus in a high dimensional space capturing many of its nonlinear features but then use a simple weighted sum of these features to classify the pattern into a few categories.”
Exploratory adaptation in large random networks
Schreier, H. I., Soen, Y., & Brenner, N. (2017) · Nature Communications · doi:10.1038/ncomms14826 · Cited by 53
Modelling Peri-Perceptual Brain Processes in a Deep Learning Spiking Neural Network Architecture
Doborjeh, Z., Kasabov, N. K., Doborjeh, M., & Sumich, A. L. (2018) · Scientific Reports · doi:10.1038/s41598-018-27169-8 · Cited by 58
“SNNs are the third generation of neural networks and comparing to conventional neuronal networks which deal with static vector-based data (temporal information needs to be converted into vectors of static features) 33 , SNNs incorporate spatio and temporal components of data into operating.”
BackgroundPerturbing low dimensional activity manifolds in spiking neuronal networks
Wärnberg, E., & Kumar, A. (2019) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1007074 · Cited by 52
“First, we consider the FORCE learning rule [12] which builds upon the liquid state machine and echo state networks [11, 17, 18].”
MethodComparison of Photonic Reservoir Computing Systems for Fiber Transmission Equalization
Argyris, A., Cantero, J., Galletero, M., Pereda, E., Mirasso, C. R., Fischer, I., & Soriano, M. C. (2019) · IEEE Journal of Selected Topics in Quantum Electronics · doi:10.1109/jstqe.2019.2936947 · Cited by 56
“2936947 is solely performed on the weights of the readout layer, while those of the recurrent network are kept fixed [2], [3].”
BackgroundDettori, S., Matino, I., Colla, V., & Speets, R. (2021) · Neural Computing and Applications · doi:10.1007/s00521-021-05984-x · Cited by 52
“Reservoir computing has been introduced by Maass et al. [9] by introducing a RNN architecture called liquid-state machine .”
MethodMovement prediction from real-world images using a liquid state machine
Burgsteiner, H., Kröll, M., Leopold, A., & Steinbauer, G. (2006) · Applied Intelligence · doi:10.1007/s10489-006-0007-1 · Cited by 63
“The LSM from [18] is the theoretical framework for computations in neural microcircuits.”
BackgroundMethodInfluentialWiest, M., Bentley, N., & Nicolelis, M. A. L. (2004) · Journal of Neurophysiology · doi:10.1152/jn.00556.2004 · Cited by 52
Resonance or Integration? Self-Sustained Dynamics and Excitability of Neural Microcircuits
Muresan, R. C., & Savin, C. (2006) · Journal of Neurophysiology · doi:10.1152/jn.01043.2006 · Cited by 54
“We investigated large networks of IF, RS, and RES neurons also referred to as " neural microcircuits " (Grillner et al. 2005; Maass et al. 2002).”
BackgroundNonlinear Modeling of Causal Interrelationships in Neuronal Ensembles
Zanos, T. P., Courellis, S. H., Berger, T. W., Hampson, R. E., Deadwyler, S. A., & Marmarelis, V. Z. (2008) · IEEE Transactions on Neural Systems and Rehabilitation Engineering · doi:10.1109/tnsre.2008.926716 · Cited by 52
Douglas, Z. H., Maniscalco, B., Hallett, M., Wassermann, E. M., & He, B. J. (2015) · The Journal of Neuroscience · doi:10.1523/jneurosci.4894-14.2015 · Cited by 51
“Mechanistically, such temporal averaging could potentially be performed by trajectorybased processing that is time sensitive (Maass et al., 2002; Rabinovich et al., 2008; He, 2013).”
BackgroundParameterizing echo state networks for multi-step time series prediction
Viehweg, J., Worthmann, K., & Mäder, P. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.11.044 · Cited by 60
Perspective on unconventional computing using magnetic skyrmions
Lee, O., Msiska, R., Brems, M. A., Kläui, M., Kurebayashi, H., & Everschor‐Sitte, K. (2023) · Applied Physics Letters · doi:10.1063/5.0148469 · Cited by 58
Streamlined photonic reservoir computer with augmented memory capabilities
Zhou, C., Huang, Y., Yang, Y., Cai, D., Zhou, P., Lau, K. Y., Li, N., & Li, X. (2024) · Opto-Electronic Advances · doi:10.29026/oea.2025.240135 · Cited by 51
Rarely categorical, highly separable representations along the cortical hierarchy
Posani, L., Wang, S., Muscinelli, S. P., Paninski, L., & Fusi, S. (2026) · Nature · doi:10.1038/s41586-026-10668-4 · Cited by 52
NeoN: Neuromorphic control for autonomous robotic navigation
Mitchell, J. P., Bruer, G. R., Dean, M. E., Plank, J. S., Rose, G. S., & Schuman, C. D. (2017) · IEEE International Symposium on Robotics and Intelligent Sensors (IRIS) · doi:10.1109/iris.2017.8250111 · Cited by 51
“Another computational model for SNNs called liquid state machines (LSM) has been developed by Maass and Markham [20].”
Separation of chaotic signals by reservoir computing
Krishnagopal, S., Girvan, M., Ott, E., & Hunt, B. R. (2020) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/1.5132766 · Cited by 51
“Our approach is based on a recurrent machine learning architecture originally known as Echo State Networks (originally proposed in the field of machine learning) [11] and Liquid State Machines (originally proposed in the field of computational neuroscience) [20], but now commonly referred to as Reservoir Computing [19…”
MethodSIMPEL: Circuit model for photonic spike processing laser neurons
Shastri, B. J., Nahmias, M. A., Tait, A. N., Wu, B., & Prucnal, P. R. (2015) · Optics Express · doi:10.1364/oe.23.008029 · Cited by 51
On Computational Power and the Order-Chaos Phase Transition in Reservoir Computing
Schrauwen, B., Buesing, L., & Legenstein, R. A. (2008) · Ghent University Academic Bibliography (Ghent University) · Cited by 66
Evolving spike-timing-dependent plasticity for single-trial learning in robots
Paolo, E. A. D. (2003) · Philosophical Transactions of the Royal Society A Mathematical Physical and Engineering Sciences · doi:10.1098/rsta.2003.1256 · Cited by 51
“They can perform novel kinds of computation, such as the recognition of temporal patterns using transient synchrony (Hop ̄eld & Brody 2001) and real-time computation without stable states in high-dimensional `liquids’ of transient activity (Maass et al. 2002).”
BackgroundMeta-Learning through Hebbian Plasticity in Random Networks
Najarro, E., & Risi, S. (2020) · IT University Of Copenhagen (IT University of Copenhagen) · doi:10.48550/arxiv.2007.02686 · Cited by 97
“For example, random connectivity seems to play a critical role in the prefrontal cortex [38], allowing an increase in the dimensionality of neural representations.”
BackgroundEmergence of a resonance in machine learning
Zhai, Z.-M., Kong, L.-W., & Lai, Y. (2023) · Physical Review Research · doi:10.1103/physrevresearch.5.033127 · Cited by 60
Error-Backpropagation in Networks of Fractionally Predictive Spiking Neurons
Bohté, S. M. (2011) · Lecture notes in computer science · doi:10.1007/978-3-642-21735-7_8 · Cited by 50
“Recurrent spiking neural network approaches can achieve longer memory, though they are notoriously hard to train [21]; edge-of-stability dynamical systems methods like reservoir computing show promise [11, 4], although they require many neurons and spikes, and mostly disregard notions of spike-time coding.”
BackgroundEffects of synaptic connectivity on liquid state machine performance
Ju, H., Xu, J., Chong, E., & VanDongen, A. M. J. (2012) · Neural Networks · doi:10.1016/j.neunet.2012.11.003 · Cited by 51
“9 × 10−8 (IE, II) (Maass et al., 2002); Wscale is a scaling factor, which is one of the parameters that we will investigate in this paper.”
BackgroundMethodInfluentialSelection of cortical dynamics for motor behaviour by the basal ganglia
Mannella, F., & Baldassarre, G. (2015) · Biological Cybernetics · doi:10.1007/s00422-015-0662-6 · Cited by 48
Recurrent Spiking Neural Network Learning Based on a Competitive Maximization of Neuronal Activity
Demin, V. A., & Nekhaev, D. (2018) · Frontiers in Neuroinformatics · doi:10.3389/fninf.2018.00079 · Cited by 51
“Liquid State Machines were also used with various kinds of plasticity, including STDP, so the model showed better performance in memorizing spatio-temporal patterns than the Liquid State Machine with static weights (Lazar et al., 2009).”
BackgroundSearching for principles of brain computation
Maass, W. (2016) · Current Opinion in Behavioral Sciences · doi:10.1016/j.cobeha.2016.06.003 · Cited by 60
Self-citationDiscrete-Time Signatures and Randomness in Reservoir Computing
Cuchiero, C., Gonon, L., Grigoryeva, L., Ortega, J., & Teichmann, J. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2021.3076777 · Cited by 50
Metadata
- Authors
- Maass, W., Natschläger, T., & Markram, H.
- Affiliations
- Graz University of Technology (AT)
- École Polytechnique Fédérale de Lausanne (CH)
- Corresponding author
- Maass, Wolfgang
- Natschläger, Thomas
- Markram, Henry
- Venue
- Neural Computation
- Publisher
- MIT Press
- Published
- 2002-11-01
- Volume
- 14
- Issue
- 11
- Pages
- 2531-2560
- ISSN
- 0899-7667 (ISSN-L)
- 1530-888X
- OpenAlex
- W2103179919
- Cited by
- 4,025
- References
- 28
- Citations by year
- 2026: 286 · 2025: 351 · 2024: 307 · 2023: 309 · 2022: 276 · 2021: 303 · 2020: 239 · 2019: 244 · 2018: 217 · 2017: 173 · 2016: 155 · 2015: 135 · 2014: 142 · 2013: 139 · 2012: 125
- FWCI
- 18.37
- Type
- journal article
- Language
- en
- Open access
- green
- infoscience.epfl.ch
- Locations
- Neural Computation (publishedVersion)
- PubMed (publishedVersion)
- Infoscience (Ecole Polytechnique Fédérale de Lausanne) (submittedVersion)
- Infoscience (Ecole Polytechnique Fédérale de Lausanne) (submittedVersion, cc-by-nc-nd)
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › Neural dynamics and brain function
- Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Reservoir Computing
- Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
- Keywords
- liquid state machines, recurrent neural circuits, transient dynamics, high-dimensional dynamical systems, fading memory, integrate-and-fire neurons, real-time processing, time-varying inputs, neural coding, statistical learning theory, computational power, attractor networks
- MeSH terms
- Action Potentials, Computer Simulation, Computer Systems, Computers, Models, Neurological, Neurons, Neural Networks, Computer
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.