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 101–200 of 1,014
Cheng, S., Quilodrán-Casas, C., Ouala, S., Farchi, A., Liu, C., Tandeo, P., Fablet, R., Lucor, D., Iooss, B., Brajard, J., et al. (2023) · IEEE/CAA Journal of Automatica Sinica · doi:10.1109/jas.2023.123537 · Cited by 303
“For instance, reservoir computing approaches [247], [248] were shown to be well-suited for learning dynamical systems from data [249], [250].”
BackgroundMemory traces in dynamical systems
Ganguli, S., Huh, D., & Sompolinsky, H. (2008) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.0804451105 · Cited by 303
On the computational power of circuits of spiking neurons
Maass, W., & Markram, H. (2004) · Journal of Computer and System Sciences · doi:10.1016/j.jcss.2004.04.001 · Cited by 301
“In this article we explore the computational power of generic circuits of spiking neurons in the light of a recently proposed new model, the liquid state machine [24].”
BackgroundMethodResultInfluentialSelf-citationDendritic spines and distributed circuits
Yuste, R. (2011) · Neuron · doi:10.1016/j.neuron.2011.07.024 · Cited by 298
“…codes and functional states, such as dynamical attractors, is a cornerstone of the neural network literature (Buonomano, 2009; Hopfield, 1982; Maass et al., 2002; Sussillo and Abbott, 2009) and is a major departure from the traditional view of using receptive field responses of individual…”
Information processing in echo state networks at the edge of chaos
Boedecker, J., Obst, O., Lizier, J. T., Mayer, N. M., & Asada, M. (2012) · Theory in Biosciences · doi:10.1007/s12064-011-0146-8 · Cited by 297
Prychynenko, D., Sitte, M., Litzius, K., Krüger, B., Bourianoff, G., Kläui, M., Sinova, J., & Everschor-Sitte, K. (2018) · Physical Review Applied · doi:10.1103/physrevapplied.9.014034 · Cited by 293
A reservoir of time constants for memory traces in cortical neurons
Bernacchia, A., Seo, H., Lee, D., & Wang, X.-J. (2011) · Nature Neuroscience · doi:10.1038/nn.2752 · Cited by 292
“If a single value function (for a given stimulus/action) and a single reward (delivered at time zero) are considered, the solution of equation (13) can be approximated by an exponential response,”
BackgroundGerstner, W., Lehmann, M., Liakoni, V., Corneil, D., & Brea, J. (2018) · Frontiers in Neural Circuits · doi:10.3389/fncir.2018.00053 · Cited by 290
“For example, memory traces hidden in the rich firing activity patterns of a recurrent network (Maass et al., 2002; Jaeger and Haas, 2004; Buonomano and Maass, 2009; Sussillo, 2009) or short-term synaptic plasticity in recurrent networks (Mongillo et al.”
BackgroundSynaptic Plasticity Dynamics for Deep Continuous Local Learning (DECOLLE)
Kaiser, J., Mostafa, H., & Neftci, E. (2020) · Frontiers in Neuroscience · doi:10.3389/fnins.2020.00424 · Cited by 288
“This is achieved using a local rate-based cost function reminiscent of readout mechanisms used in liquid state machines [15], but where the readout is performed over a fixed random combination of the neuron outputs.”
MethodRecurrent Neural Networks for Short-Term Load Forecasting: An Overview and Comparative Analysis
Bianchi, F. M., Maiorino, E., Kampffmeyer, M. C., Rizzi, A., & Jenssen, R. (2017) · SpringerBriefs in Computer Science · doi:10.1007/978-3-319-70338-1 · Cited by 280
“ESNs, along with Liquid State Machines [137], belong to the class of computational dynamical systems implemented according to the so-called reservoir computing framework [71].”
BackgroundReservoir Computing Using Diffusive Memristors
Midya, R., Wang, Z., Asapu, S., Zhang, X., Rao, M., Song, W., Zhuo, Y., Upadhyay, N., Xia, Q., & Yang, J. J. (2019) · Advanced Intelligent Systems · doi:10.1002/aisy.201900084 · Cited by 279
Working models of working memory
Barak, O., & Tsodyks, M. (2014) · Current Opinion in Neurobiology · doi:10.1016/j.conb.2013.10.008 · Cited by 278
“An alternative view is that memory of an item could be maintained by highly non-stationary activity, as illustrated by the framework of reservoir computing [41,42].”
BackgroundToward optical signal processing using Photonic Reservoir Computing
Vandoorne, K., Dierckx, W., Schrauwen, B., Verstraeten, D., Baets, R., Bienstman, P., & Campenhout, J. V. (2008) · Optics Express · doi:10.1364/oe.16.011182 · Cited by 276
Echo state networks are universal
Grigoryeva, L., & Ortega, J.-P. (2018) · Neural Networks · doi:10.1016/j.neunet.2018.08.025 · Cited by 274
Plasticity in single neuron and circuit computations
Destexhe, A., & Marder, E. (2004) · Nature · doi:10.1038/nature03011 · Cited by 274
Model Predictive Control of Unknown Nonlinear Dynamical Systems Based on Recurrent Neural Networks
Pan, Y., & Wang, J. (2012) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2011.2169636 · Cited by 273
“A similar idea of the ESN has been discovered independently by Maass [52] for development of the " liquid state machines.”
BackgroundComputing with Spiking Neuron Networks
Paugam-Moisy, H., & Bohte, S. (2012) · Handbook of Natural Computing · doi:10.1007/978-3-540-92910-9_10 · Cited by 272
“The basic motivation of the Liquid State Machine, defined by Maass, Natschläger and Markram in 2002 [101], was to explain how a continuous stream of inputs u(.”
BackgroundA survey on the application of recurrent neural networks to statistical language modeling
De Mulder, W., Bethard, S., & Moens, M.-F. (2015) · Computer Speech & Language · doi:10.1016/j.csl.2014.09.005 · Cited by 272
“Description In 2001 and 2002 a fundamentally new approach to RNN design and training was proposed independently by Maass under the name of Liquid State Machines (Maass et al., 2002) and by Jaeger under the name of Echo State Networks (Jaeger, 2001).”
MethodSPAN: SPIKE PATTERN ASSOCIATION NEURON FOR LEARNING SPATIO-TEMPORAL SPIKE PATTERNS
MOHEMMED, A., SCHLIEBS, S., MATSUDA, S., & KASABOV, N. (2012) · International Journal of Neural Systems · doi:10.1142/s0129065712500128 · Cited by 270
Neural heterogeneity promotes robust learning
Perez-Nieves, N., Leung, V. C. H., Dragotti, P. L., & Goodman, D. F. M. (2021) · Nature Communications · doi:10.1038/s41467-021-26022-3 · Cited by 267
Exploring Neuromorphic Computing Based on Spiking Neural Networks: Algorithms to Hardware
Rathi, N., Chakraborty, I., Kosta, A., Sengupta, A., Ankit, A., Panda, P., & Roy, K. (2023) · ACM Computing Surveys · doi:10.1145/3571155 · Cited by 267
“Liquid state machines (LSM) have been explored as lightweight architectures to handle spatio-temporal inputs in a bio-plausible manner [134, 136].”
BackgroundLegenstein, R., Pecevski, D., & Maass, W. (2008) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1000180 · Cited by 267
“This learning theory for reward-modulated STDP suggests that the main functional role of noise is to maintain a suitable level of spontaneous firing (since if a neuron does not fire, it cannot find out whether this will be rewarded), which should vary from trial to trial in order to explore which firing patterns are r…”
BackgroundSelf-citationPredictive Coding of Dynamical Variables in Balanced Spiking Networks
Boerlin, M., Machens, C. K., & Denève, S. (2013) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1003258 · Cited by 266
“This approach differs from the ‘‘liquid computing’’ approach in which recurrent networks have fixed, random connectivities while the decoding weights are learnt [27].”
MethodHigh-performance photonic reservoir computer based on a coherently driven passive cavity
Vinckier, Q., Duport, F., Smerieri, A., Vandoorne, K., Bienstman, P., Haelterman, M., & Massar, S. (2015) · Optica · doi:10.1364/optica.2.000438 · Cited by 265
“Reservoir computing is a recent bio-inspired method for processing information [1-3] that uses a recurrent dynamical system called “reservoir” to process time-dependent signals.”
MethodSelective corticostriatal plasticity during acquisition of an auditory discrimination task
Xiong, Q., Znamenskiy, P., & Zador, A. M. (2015) · Nature · doi:10.1038/nature14225 · Cited by 263
Pattern Recognition in a Bucket
Fernando, C., & Sojakka, S. (2003) · Lecture Notes in Computer Science · doi:10.1007/978-3-540-39432-7_63 · Cited by 262
“Maass et al [1] have produced a model called the Liquid State Machine (LSM) which hypothesises that a cortical microcolumn consists of stereotyped recurrent circuits of integrate-and-fire neurons connected randomly according to a few parameters.”
BackgroundIsolated word recognition with the Liquid State Machine: a case study
Verstraeten, D., Schrauwen, B., Stroobandt, D., & Van Campenhout, J. (2005) · Information Processing Letters · doi:10.1016/j.ipl.2005.05.019 · Cited by 262
Recent progress in semiconductor excitable lasers for photonic spike processing
Prucnal, P. R., Shastri, B. J., Ferreira de Lima, T., Nahmias, M. A., & Tait, A. N. (2016) · Advances in Optics and Photonics · doi:10.1364/aop.8.000228 · Cited by 261
“This distributed, asynchronous model processes information using both space and time [56,57].”
MethodAgarap, A. F. M. (2018) · Proceedings of the 2018 10th International Conference on Machine Learning and Computing · doi:10.1145/3195106.3195117 · Cited by 259
“ization algorithm (training algorithm), uncertainty, feature extraction, nor network traffic data collection. 2 LITERATURE REVIEW The ESN is a type of reservoir computing (RC) proposed by Maass (2002)[35], which uses the output of the network to train a simple linear read-out function. The read-out function is simply…”
BackgroundHow critical is brain criticality?
O’Byrne, J., & Jerbi, K. (2022) · Trends in Neurosciences · doi:10.1016/j.tins.2022.08.007 · Cited by 258
Ganguli, S., & Sompolinsky, H. (2012) · Annual Review of Neuroscience · doi:10.1146/annurev-neuro-062111-150410 · Cited by 256
“More recent proposals ( Jaeger 2001, Maass et al. 2002, Jaeger & Haas 2004) suggest that recurrent networks could store temporal sequences of inputs in their ongoing, transient activity.”
BackgroundA theoretical and experimental study of neuromorphic atomic switch networks for reservoir computing
Sillin, H. O., Aguilera, R., Shieh, H.-H., Avizienis, A. V., Aono, M., Stieg, A. Z., & Gimzewski, J. K. (2013) · Nanotechnology · doi:10.1088/0957-4484/24/38/384004 · Cited by 254
High-Performance OCR for Printed English and Fraktur Using LSTM Networks
Breuel, T. M., Ul-Hasan, A., Al-Azawi, M. A., & Shafait, F. (2013) · 2013 12th International Conference on Document Analysis and Recognition · doi:10.1109/icdar.2013.140 · Cited by 250
“The LSTM architecture differs significantly from the architecture of previous recurrent neural networks [11], [12], [13] and appears to overcome many of the limitations and problems of those earlier architectures.”
BackgroundNeural circuits as computational dynamical systems
Sussillo, D. (2014) · Current Opinion in Neurobiology · doi:10.1016/j.conb.2014.01.008 · Cited by 247
“training just W, known as ‘reservoir computing’) [30,34–36], or to improved optimization algorithms [37,38 ,39–41], RNNs are enjoying a renaissance.”
History-dependent variability in population dynamics during evidence accumulation in cortex
Morcos, A. S., & Harvey, C. D. (2016) · Nature Neuroscience · doi:10.1038/nn.4403 · Cited by 247
“This new model is consistent with the theoretical framework developed in reservoir computing 45,48,49 .”
Recurrent neural networks as versatile tools of neuroscience research
Barak, O. (2017) · Current Opinion in Neurobiology · doi:10.1016/j.conb.2017.06.003 · Cited by 244
Laser dynamical reservoir computing with consistency: an approach of a chaos mask signal
Nakayama, J., Kanno, K., & Uchida, A. (2016) · Optics Express · doi:10.1364/oe.24.008679 · Cited by 244
Data-driven science and engineering : machine learning, dynamical systems, and control
Brunton, S. L., & Kutz, J. N. (2019) · doi:10.1017/978-1-009-08951-7 · Cited by 242
Computational Aspects of Feedback in Neural Circuits
Maass, W., Joshi, P., & Sontag, E. D. (2007) · PLoS Computational Biology · doi:10.1371/journal.pcbi.0020165 · Cited by 242
“, they are input-driven and have no ‘‘internal clock’’) and have a fading memory (see [5]).”
BackgroundMethodInfluentialSelf-citationA survey on predicting the popularity of web content
Tatar, A., de Amorim, M. D., Fdida, S., & Antoniadis, P. (2014) · Journal of Internet Services and Applications · doi:10.1186/s13174-014-0008-y · Cited by 240
“Reservoir computing [76], a novel paradigm in recurrent neural networks, is proposed as a model that could consider more complex interactions between early and late popularity values (between Xc(ti) and Nc(tr)).”
BackgroundReservoir Computing Beyond Memory-Nonlinearity Trade-off
Inubushi, M., & Yoshimura, K. (2017) · Scientific Reports · doi:10.1038/s41598-017-10257-6 · Cited by 239
“This framework is referred to as reservoir computing ( RC ) 5 which was proposed originally in the research fields of machine learning 6 (called Echo State Network) and computational neuroscience 7 (called Liquid State Machine).”
BackgroundPhotonic multiplexing techniques for neuromorphic computing
Bai, Y., Xu, X., Tan, M., Sun, Y., Li, Y., Wu, J., Morandotti, R., Mitchell, A., Xu, K., & Moss, D. J. (2023) · Nanophotonics · doi:10.1515/nanoph-2022-0485 · Cited by 235
The BrainScaleS-2 Accelerated Neuromorphic System With Hybrid Plasticity
Pehle, C., Billaudelle, S., Cramer, B., Kaiser, J., Schreiber, K., Stradmann, Y., Weis, J., Leibfried, A., Müller, E., & Schemmel, J. (2022) · Frontiers in Neuroscience · doi:10.3389/fnins.2022.795876 · Cited by 233
“These dynamics can be deliberately exploited for information processing by tying on the reservoir computing framework [34, 35].”
BackgroundThe neuroconnectionist research programme
Doerig, A., Sommers, R. P., Seeliger, K., Richards, B., Ismael, J., Lindsay, G. W., Kording, K. P., Konkle, T., van Gerven, M. A. J., Kriegeskorte, N., et al. (2023) · Nature Reviews Neuroscience · doi:10.1038/s41583-023-00705-w · Cited by 233
The Non-Random Brain: Efficiency, Economy, and Complex Dynamics
Sporns, O. (2011) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2011.00005 · Cited by 233
“An attractive hypothesis suggests that dynamics in local circuits gives rise to a “computational reservoir” that can be molded by synaptic plasticity and support a variety of functional outcomes (Maass et al., 2002).”
BackgroundReservoir Computing With Spin Waves Excited in a Garnet Film
Nakane, R., Tanaka, G., & Hirose, A. (2018) · IEEE Access · doi:10.1109/access.2018.2794584 · Cited by 231
Inhibitory Plasticity: Balance, Control, and Codependence
Hennequin, G., Agnes, E. J., & Vogels, T. P. (2017) · Annual Review of Neuroscience · doi:10.1146/annurev-neuro-072116-031005 · Cited by 228
“Recent research on dynamical systems has revisited this idea, proposing that neuronal networks represent memories through the dynamic paths traced out by the activity of the whole population (Maass et al. 2002, Sussillo & Abbott 2009, Laje & Buonomano 2013, Hennequin et al. 2014, Song et al. 2016).”
BackgroundImproving reservoirs using intrinsic plasticity
Schrauwen, B., Wardermann, M., Verstraeten, D., Steil, J. J., & Stroobandt, D. (2008) · Neurocomputing · doi:10.1016/j.neucom.2007.12.020 · Cited by 227
“The LSM by Maass [20] was originally presented as a general framework to perform real-time computation on temporal signals, but most of its realizations are based on an abstract cortical micro-column model.”
BackgroundMethodPhysical reservoir computing with emerging electronics
Liang, X., Tang, J., Zhong, Y., Gao, B., Qian, H., & Wu, H. (2024) · Nature Electronics · doi:10.1038/s41928-024-01133-z · Cited by 226
“In 2002, Wolfgang Maass proposed the liquid-state machine (LSM), which operates in a continuous and spiking fashion 14,15 .”
Reservoir Computing: Theory, Physical Implementations, and Applications
(2021) · Natural Computing Series · doi:10.1007/978-981-13-1687-6 · Cited by 225
Recent progress in analog memory-based accelerators for deep learning
Tsai, H., Ambrogio, S., Narayanan, P., Shelby, R. M., & Burr, G. W. (2018) · Journal of Physics D: Applied Physics · doi:10.1088/1361-6463/aac8a5 · Cited by 225
Hierarchical Neural Networks for Image Interpretation
Behnke, S. (2003) · Lecture Notes in Computer Science · doi:10.1007/b11963 · Cited by 225
Reservoir Computing Approaches for Representation and Classification of Multivariate Time Series
Bianchi, F. M., Scardapane, S., Lokse, S., & Jenssen, R. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.3001377 · Cited by 224
“For this reason, several authors proposed to replace the linear decoding function g(·) in (5) with a nonlinear model, such as SVMs [12] or MLPs [41]–[43].”
BackgroundInfluentialParallel Reservoir Computing Using Optical Amplifiers
Vandoorne, K., Dambre, J., Verstraeten, D., Schrauwen, B., & Bienstman, P. (2011) · IEEE Transactions on Neural Networks · doi:10.1109/tnn.2011.2161771 · Cited by 224
“I. INTRODUCTION RESERVOIR computing (RC) is a training concept forrecurrent neural networks (RNNs), introduced a few years ago, that combines the advantages of both recurrent and feed forward networks [1], [2].”
BackgroundTransition to Chaos in Random Neuronal Networks
Kadmon, J., & Sompolinsky, H. (2015) · Physical Review X · doi:10.1103/physrevx.5.041030 · Cited by 283
Manjunath, G., & Jaeger, H. (2013) · Neural Computation · doi:10.1162/neco_a_00411 · Cited by 222
“ESNs are artificial recurrent neural networks (RNNs) which are used in machine learning for the supervised training of temporal pattern recognizers, pattern generators, predictors, controllers and more (short overview: [12]; target paper: [10]; survey on state of the art: [16]; other RC flavors besides ESNs: liquid st…”
BackgroundAutomated EEG-Based Diagnosis of Neurological Disorders: Inventing the Future of Neurology
Adeli, H., & Ghosh‐Dastidar, S. (2010) · CRC Press eBooks · doi:10.1201/9781439815328 · Cited by 244
Haeusler, S., & Maass, W. (2006) · Cerebral Cortex · doi:10.1093/cercor/bhj132 · Cited by 221
Self-citationVedmedenko, E. Y., Kawakami, R. K., Sheka, D. D., Gambardella, P., Kirilyuk, A., Hirohata, A., Binek, C., Chubykalo-Fesenko, O., Sanvito, S., Kirby, B. J., et al. (2020) · Journal of Physics D: Applied Physics · doi:10.1088/1361-6463/ab9d98 · Cited by 220
Emergent Criticality in Complex Turing B‐Type Atomic Switch Networks
Stieg, A. Z., Avizienis, A. V., Sillin, H. O., Martin‐Olmos, C., Aono, M., & Gimzewski, J. K. (2012) · Advanced Materials · doi:10.1002/adma.201103053 · Cited by 217
Novel frontier of photonics for data processing—Photonic accelerator
Kitayama, K.-I., Notomi, M., Naruse, M., Inoue, K., Kawakami, S., & Uchida, A. (2019) · APL Photonics · doi:10.1063/1.5108912 · Cited by 216
“34,35 The training process of the connection weights is simplified in reservoir computing, and only the output weights are trained by learning.”
BackgroundVariability in neuronal activity in primate cortex during working memory tasks
Shafi, M., Zhou, Y., Quintana, J., Chow, C., Fuster, J., & Bodner, M. (2007) · Neuroscience · doi:10.1016/j.neuroscience.2006.12.072 · Cited by 215
Computational principles of memory
Chaudhuri, R., & Fiete, I. (2016) · Nature Neuroscience · doi:10.1038/nn.4237 · Cited by 213
Hramov, A. E., Maksimenko, V. A., & Pisarchik, A. N. (2021) · Physics Reports · doi:10.1016/j.physrep.2021.03.002 · Cited by 213
Singer, W. (2013) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2013.09.006 · Cited by 212
Supervised learning in spiking neural networks with FORCE training
Nicola, W., & Clopath, C. (2017) · Nature Communications · doi:10.1038/s41467-017-01827-3 · Cited by 211
SORN: a Self-organizing Recurrent Neural Network
Lazar, A. (2009) · Frontiers in Computational Neuroscience · doi:10.3389/neuro.10.023.2009 · Cited by 209
“Such static networks have received signifi cant attention in the recent past, giving rise to the fi eld of reservoir computing (Jaeger, 2001; Maass et al., 2002).”
The Sparseness of Mixed Selectivity Neurons Controls the Generalization–Discrimination Trade-Off
Barak, O., Rigotti, M., & Fusi, S. (2013) · The Journal of Neuroscience · doi:10.1523/jneurosci.2753-12.2013 · Cited by 207
Fully analogue photonic reservoir computer
Duport, F., Smerieri, A., Akrout, A., Haelterman, M., & Massar, S. (2016) · Scientific Reports · doi:10.1038/srep22381 · Cited by 206
“Note that since the impulse response lasts longer than T’, the current output y(n) contains contributions from the light intensity x(t) up to l input periods in the past, which is a small difference with respect to the traditional reservoir computer, see equation (2).”
BackgroundPhotonic Nonlinear Transient Computing with Multiple-Delay Wavelength Dynamics
Martinenghi, R., Rybalko, S., Jacquot, M., Chembo, Y. K., & Larger, L. (2012) · Physical Review Letters · doi:10.1103/physrevlett.108.244101 · Cited by 205
Building functional networks of spiking model neurons
Abbott, L. F., DePasquale, B., & Memmesheimer, R.-M. (2016) · Nature Neuroscience · doi:10.1038/nn.4241 · Cited by 204
“…time 1,2,3,4,5,6,7,8,9,10,11,12 , responding to particular temporal input sequences 13,14,15 , responding after a delay or activity sequence 13,16,17,18,19,20,21,22,23 , responding with a temporally complex output 7,24,25,26,27,28,29,22,23 , or autonomously generating complex dynamics…”
BackgroundThe chronotron: a neuron that learns to fire temporally-precise spike patterns
Florian, R. (2010) · Nature Precedings · doi:10.1038/npre.2010.5190.1 · Cited by 204
“This problem is very similar to the problem that needs to be solved in reservoir computing (Jaeger, 2001; Maass et al., 2002; Schrauwen et al., 2007), where the state of a high-dimensional dynamical system, such as our vector λ, is processed by a (usually) linear discriminator such that the switch between output state…”
BackgroundMethodScalable reservoir computing on coherent linear photonic processor
Nakajima, M., Tanaka, K., & Hashimoto, T. (2021) · Communications Physics · doi:10.1038/s42005-021-00519-1 · Cited by 203
Distributed Fading Memory for Stimulus Properties in the Primary Visual Cortex
Nikolić, D., Häusler, S., Singer, W., & Maass, W. (2009) · PLoS Biology · doi:10.1371/journal.pbio.1000260 · Cited by 202
“This suggests that the brain exploits the computational advantages of such processes as predicted by computer simulations of generic cortical microcircuits [31,62], not only for high-level computations, but with all likelihood at all levels of cortical processing.”
BackgroundSelf-citationZylberberg, J., & Strowbridge, B. W. (2017) · Annual Review of Neuroscience · doi:10.1146/annurev-neuro-070815-014006 · Cited by 202
“…balance between excitation and inhibition in the circuit, information about the network’s inputs will be maintained after those inputs go away (Bertschinger & Natschläger 2004, Buonomano & Maass 2009, Laje & Buonomano 2013, Maass et al. 2002, Singh & Eliasmith 2006, Toyoizumi & Abbott 2011).”
BackgroundLiquid State Machines: Motivation, Theory, and Applications
Maass, W. (2011) · Computability in Context · doi:10.1142/9781848162778_0008 · Cited by 201
“4, which had been introduced in [1]:”
BackgroundMethodSelf-citationA Flexible Model of Working Memory
Bouchacourt, F., & Buschman, T. J. (2019) · Neuron · doi:10.1016/j.neuron.2019.04.020 · Cited by 201
Hands-on reservoir computing: a tutorial for practical implementation
Cucchi, M., Abreu, S., Ciccone, G., Brunner, D., & Kleemann, H. (2022) · Neuromorphic Computing and Engineering · doi:10.1088/2634-4386/ac7db7 · Cited by 199
A Unified Mathematical Framework for Coding Time, Space, and Sequences in the Hippocampal Region
Howard, M. W., MacDonald, C. J., Tiganj, Z., Shankar, K. H., Du, Q., Hasselmo, M. E., & Eichenbaum, H. (2014) · The Journal of Neuroscience · doi:10.1523/jneurosci.5808-12.2014 · Cited by 199
“In the temporal domain, the present mathematical framework is a special case of a liquid state machine (Maass et al., 2002; Buonomano and Maass, 2009), although an extremely simple one that is trivially decodable.”
BackgroundOptoelectronic oscillators with time-delayed feedback
Chembo, Y. K., Brunner, D., Jacquot, M., & Larger, L. (2019) · Reviews of Modern Physics · doi:10.1103/revmodphys.91.035006 · Cited by 199
“In order to overcome this difficulty, Maass, Natschläger, and Markram (2002) introduced the liquid-state machine (LSM), while Jaeger and Haas (2004) independently proposed the idea of echo state network (ESN).”
BackgroundArchitectural and Markovian factors of echo state networks
Gallicchio, C., & Micheli, A. (2011) · Neural Networks · doi:10.1016/j.neunet.2011.02.002 · Cited by 198
Nonlocal Nonlinear Electro-Optic Phase Dynamics Demonstrating 10 Gb/s Chaos Communications
Lavrov, R., Jacquot, M., & Larger, L. (2010) · IEEE Journal of Quantum Electronics · doi:10.1109/jqe.2010.2049987 · Cited by 204
RECURRENT NEURAL NETWORKS ARE UNIVERSAL APPROXIMATORS
SCHÄFER, A. M., & ZIMMERMANN, H.-G. (2007) · International Journal of Neural Systems · doi:10.1142/s0129065707001111 · Cited by 196
Why Brain Criticality Is Clinically Relevant: A Scoping Review
Zimmern, V. (2020) · Frontiers in Neural Circuits · doi:10.3389/fncir.2020.00054 · Cited by 194
Encoding and Decoding of Overlapping Odor Sequences
Broome, B. M., Jayaraman, V., & Laurent, G. (2006) · Neuron · doi:10.1016/j.neuron.2006.07.018 · Cited by 192
“How randomly connected recurrent networks might process time-varying continuous input is the subject of recent theoretical work on ‘‘liquid state machines’’ (Maass et al., 2002).”
All-optical reservoir computer based on saturation of absorption
Dejonckheere, A., Duport, F., Smerieri, A., Fang, L., Oudar, J.-L., Haelterman, M., & Massar, S. (2014) · Optics Express · doi:10.1364/oe.22.010868 · Cited by 189
Reservoir Computing Properties of Neural Dynamics in Prefrontal Cortex
Enel, P., Procyk, E., Quilodran, R., & Dominey, P. F. (2016) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1004967 · Cited by 187
Hao, Y., Huang, X., Dong, M., & Xu, B. (2020) · Neural Networks · doi:10.1016/j.neunet.2019.09.007 · Cited by 187
“The LSM is a classic model of reservoir computing (RC) [67, 68, 69, 70, 64], which”
BackgroundMethodOptoelectronic reservoir computing: tackling noise-induced performance degradation
Soriano, M. C., Ortín, S., Brunner, D., Larger, L., Mirasso, C. R., Fischer, I., & Pesquera, L. (2013) · Optics Express · doi:10.1364/oe.21.000012 · Cited by 187
Histed, M. H., Pasupathy, A., & Miller, E. K. (2009) · Neuron · doi:10.1016/j.neuron.2009.06.019 · Cited by 187
“Demonstrating the feasibility of both methods, neural network models have been devised that use each method for information storage (Hopfield, 1982; Rumelhart et al., 1986; Maass et al., 2002; Drew and Abbott, 2006; Ganguli et al., 2008).”
Large-Scale Optical Reservoir Computing for Spatiotemporal Chaotic Systems Prediction
Rafayelyan, M., Dong, J., Tan, Y., Krzakala, F., & Gigan, S. (2020) · Physical Review X · doi:10.1103/physrevx.10.041037 · Cited by 186
What Is Morphological Computation? On How the Body Contributes to Cognition and Control
Müller, V. C., & Hoffmann, M. (2017) · Artificial Life · doi:10.1162/artl_a_00219 · Cited by 186
“[49] present a computational model based on transient dynamics in high-dimensional neural circuits.”
BackgroundPhotonic machine learning implementation for signal recovery in optical communications
Argyris, A., Bueno, J., & Fischer, I. (2018) · Scientific Reports · doi:10.1038/s41598-018-26927-y · Cited by 186
“The originally proposed RC concept 11,12 for data processing consists of three layers: the input, the reservoir and the output layer.”
Relative Refractory Period in an Excitable Semiconductor Laser
Selmi, F., Braive, R., Beaudoin, G., Sagnes, I., Kuszelewicz, R., & Barbay, S. (2014) · Physical Review Letters · doi:10.1103/physrevlett.112.183902 · Cited by 186
Lansner, A. (2009) · Trends in Neurosciences · doi:10.1016/j.tins.2008.12.002 · Cited by 186
Decoupled echo state networks with lateral inhibition
Xue, Y., Yang, L., & Haykin, S. (2007) · Neural Networks · doi:10.1016/j.neunet.2007.04.014 · Cited by 186
Six Networks on a Universal Neuromorphic Computing Substrate
Pfeil, T., Grübl, A., Jeltsch, S., Müller, E., Müller, P., Petrovici, M. A., Schmuker, M., Brüderle, D., Schemmel, J., & Meier, K. (2013) · Frontiers in Neuroscience · doi:10.3389/fnins.2013.00011 · Cited by 185
“Here, we have chosen a simple binary task adapted from Maass et al. (2002), to evaluate the performance of the LSM.”
BackgroundMethodInfluentialRemaining useful life prediction of PEMFC systems based on the multi-input echo state network
Hua, Z., Zheng, Z., Péra, M.-C., & Gao, F. (2020) · Applied Energy · doi:10.1016/j.apenergy.2020.114791 · Cited by 185
Rigotti, M. (2010) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2010.00024 · Cited by 184
“Generalizing to different timings is a problem for alternative approaches that encode all the detailed time information (Maass et al., 2002, 2007; Jaeger and Haas, 2004) or for feed-forward models of working memory (Goldman, 2009).”
BackgroundMethodThe cerebellum as a liquid state machine
Yamazaki, T., & Tanaka, S. (2007) · Neural Networks · doi:10.1016/j.neunet.2007.04.004 · Cited by 184
“The organization of the model cerebellar cortex looks similar to that of a liquid state machine (Maass et al., 2002).”
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.