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 901–1,000 of 1,014
A Coupled Spintronics Neuromorphic Approach for High‐Performance Reservoir Computing
Akashi, N., Kuniyoshi, Y., Tsunegi, S., Taniguchi, T., Nishida, M., Sakurai, R., Wakao, Y., Kawashima, K., & Nakajima, K. (2022) · Advanced Intelligent Systems · doi:10.1002/aisy.202200123 · Cited by 37
Computational capability of liquid state machines with spike-timing-dependent plasticity
Xue, F., Hou, Z., & Li, X. (2013) · Neurocomputing · doi:10.1016/j.neucom.2013.06.019 · Cited by 40
“Instead of online updating all of the synaptic conductivities in most of the traditional recurrent neural networks (RNN), synaptic weights of recurrent connections in LC are usually chosen randomly and fixed during the training process; while only the weights from neurons in LC to neurons in RC are trained by learning…”
BackgroundMethodInfluentialBrain-Inspired Classroom Occupancy Monitoring on a Low-Power Mobile Platform
Conti, F., Pullini, A., & Benini, L. (2014) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2014.95 · Cited by 42
“Important academic research lines focusing on computation (as opposed to the simulation of the brain) include liquid state machines [20] and reservoir computing [23][22].”
BackgroundBeyond Scale-Free Small-World Networks: Cortical Columns for Quick Brains
Stoop, R., Saase, V., Wagner, C., Stoop, B., & Stoop, R. (2013) · Physical Review Letters · doi:10.1103/physrevlett.110.108105 · Cited by 41
Sequence learning with hidden units in spiking neural networks
Brea, J., Senn, W., & Pfister, J.-P. (2011) · Bern Open Repository and Information System (University of Bern) · Cited by 37
wyffels, F., Schrauwen, B., Verstraeten, D., & Stroobandt, D. (2008) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2008.4634252 · Cited by 47
“networks of analog (in which case they are referred to as Echo State Networks [1]) or spiking neurons (then they are called Liquid State Machines [2]).”
BackgroundReservoir Computing with Stochastic Bitstream Neurons
Verstraeten, D., Schrauwen, B., & Stroobandt, D. (2005) · Ghent University Academic Bibliography (Ghent University) · Cited by 37
Training Passive Photonic Reservoirs With Integrated Optical Readout
Freiberger, M., Katumba, A., Bienstman, P., & Dambre, J. (2018) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2018.2874571 · Cited by 41
“Among these, Reservoir Computing [2]–[4], a machine learning technique proposed to perform brain-inspired computing, has been applied in unconventional computing recently in attempts to overcome the Von Neumann bottleneck.”
BackgroundContext-dependent selection of visuomotor maps
Salinas, E. (2004) · BMC Neuroscience · doi:10.1186/1471-2202-5-47 · Cited by 40
“It also works for calculating non-trivial functions of time using spiking neurons [38].”
BackgroundMethods for Estimating the Computational Power and Generalization Capability of Neural Microcircuits
Maass, W., Legenstein, R. A., & Bertschinger, N. (2004) · neural information processing systems · Cited by 37
Self-citationUnderstanding Networks of Computing Chemical Droplet Neurons Based on Information Flow
Gruenert, G., Giżyński, K., Escuela, G., Ibrahim, B., Górecki, J., & Dittrich, P. (2014) · International Journal of Neural Systems · doi:10.1142/s0129065714500324 · Cited by 36
Wave Turbulence and Energy Cascade in the Hippocampus
Sheremet, A., Qin, Y. F., Kennedy, J. P., Zhou, Y., & Maurer, A. P. (2019) · Frontiers in Systems Neuroscience · doi:10.3389/fnsys.2018.00062 · Cited by 40
Chen, Y., Wang, Z., Patil, A., & Basu, A. (2019) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2018.2889779 · Cited by 36
“It is broadly a part of randomized neural networks with other similar networks such as liquid state machine [19] and neural engineering framework [20] being proposed by the computaational neuroscience community.”
BackgroundOptimal short-term memory before the edge of chaos in driven random recurrent networks
Haruna, T., & Nakajima, K. (2019) · Physical review. E · doi:10.1103/physreve.100.062312 · Cited by 39
Inhibitory Plasticity: From Molecules to Computation and Beyond
Gandolfi, D., Bigiani, A., Porro, C. A., & Mapelli, J. (2020) · International Journal of Molecular Sciences · doi:10.3390/ijms21051805 · Cited by 36
Guo, X., Xiang, S., Zhang, Y. H., Lin, L., Wen, A. J., & Hao, Y. (2020) · IEEE Journal of Selected Topics in Quantum Electronics · doi:10.1109/jstqe.2020.2987077 · Cited by 40
LSMCore: A 69k-Synapse/mm 2 Single-Core Digital Neuromorphic Processor for Liquid State Machine
Wang, L., Yang, Z., Guo, S., Qu, L., Zhang, X., Kang, Z., & Xu, W. (2022) · IEEE Transactions on Circuits and Systems I Regular Papers · doi:10.1109/tcsi.2022.3147380 · Cited by 38
Dynamic Nonlinear Behavior of Ionic Liquid-Based Reservoir Computing Devices
Matsuo, T., Sato, D., Koh, S.-G., Shima, H., Naitoh, Y., Akinaga, H., Itoh, T., Nokami, T., Kobayashi, M., & Kinoshita, K. (2022) · ACS Applied Materials & Interfaces · doi:10.1021/acsami.2c04167 · Cited by 37
“The solvated IL, a composite of Cu(Tf 2 N) 2 and 2,5,8,11-tetraoxadodecane (G3) [Cu(Tf 2 N) 2 / G3 = 1:1], allowed easy access of Cu 2+ to the anode, as the ions were coordinated by electrically neutral G3 molecules.”
BackgroundInformation Processing Capacity of Spintronic Oscillator
Tsunegi, S., Kubota, T., Kamimaki, A., Grollier, J., Cros, V., Yakushiji, K., Fukushima, A., Yuasa, S., Kubota, H., Nakajima, K., et al. (2023) · Advanced Intelligent Systems · doi:10.1002/aisy.202300175 · Cited by 46
Basterrech, S., Alba, E., & Snåšel, V. (2014) · World Congress on Nature and Biologically Inspired Computing (NaBIC) · doi:10.1109/nabic.2014.6921880 · Cited by 40
Information Processing in the Hypothalamus: Peptides and Analogue Computation
Leng, G., & Ludwig, M. (2006) · Journal of Neuroendocrinology · doi:10.1111/j.1365-2826.2006.01428.x · Cited by 37
Typlt, M., Englitz, B., Sonntag, M., Dehmel, S., Kopp‐Scheinpflug, C., & Ruebsamen, R. (2012) · PLoS ONE · doi:10.1371/journal.pone.0029965 · Cited by 36
Dominey, P. F. (2013) · Frontiers in Psychology · doi:10.3389/fpsyg.2013.00500 · Cited by 36
“Subsequent work with such systems demonstrated their vast computational power (Maass et al., 2002; Jaeger and Haas, 2004).”
BackgroundReservoir computing with noise
Nathe, C., Pappu, C. S., Mecholsky, N. A., Hart, J. D., Carroll, T. L., & Sorrentino, F. (2023) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0130278 · Cited by 38
The computational power of the human brain
Gebicke‐Haerter, P. J. (2023) · Frontiers in Cellular Neuroscience · doi:10.3389/fncel.2023.1220030 · Cited by 41
“As basis sets (or “reservoirs”), randomly connected recurrent networks, like “liquid-” (Maass et al., 2002) or “echo-state machines” (Jaeger and Haas, 2004) have been constructed.”
BackgroundFry, D., Deshmukh, A. A., Chen, S. Y.-C., Rastunkov, V., & Markov, V. (2023) · Scientific Reports · doi:10.1038/s41598-023-45015-4 · Cited by 52
Resistive memory-based zero-shot liquid state machine for multimodal event data learning
Lin, N., Wang, S., Li, Y., Wang, B., Shi, S., He, Y., Zhang, W., Yu, Y., Zhang, Y., Zhang, X., et al. (2025) · Nature Computational Science · doi:10.1038/s43588-024-00751-z · Cited by 41
Dasgupta, S., Goldschmidt, D., Wörgötter, F., & Manoonpong, P. (2015) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2015.00010 · Cited by 35
“Reservoir-based recurrent neural networks (RNNs) (Maass et al., 2002; Jaeger and Haas, 2004; Sussillo and Abbott, 2009), with their inherent ability to deal with temporal information and fading memory of sensory stimuli, thus provide a suitable platform to model such internal predictive mechanisms.”
BackgroundRasch, M. J., Schuch, K., Logothetis, N. K., & Maass, W. (2010) · Journal of Neurophysiology · doi:10.1152/jn.00845.2009 · Cited by 35
“1Institute for Theoretical Computer Science, Graz University of Technology, Graz, Austria; 2Institute of Neuroscience, Chinese Academy of Science, Shanghai, China; 3Max Planck Institute for Biological Cybernetics, Tübingen, Germany; and 4Imaging Science and Biomedical”
BackgroundMethodInfluentialSelf-citationMotor-Skill Learning in an Insect Inspired Neuro-Computational Control System
Arena, E., Arena, P., Strauß, R., & Patané, L. (2017) · Frontiers in Neurorobotics · doi:10.3389/fnbot.2017.00012 · Cited by 36
“Among the different kinds of neural networks used for solving problems like navigation (Tani, 1996), multi-link system control (Cruse, 2002) and classification, a lot of interest was devoted to Reservoir computing, which mainly includes two different approaches: Echo State Network (ESN) and LSM (Jaeger, 2001; Maass et al…”
BackgroundMethodInfluentialRandomized Machine Learning Approaches: Recent Developments and Challenges
Gallicchio, C., Martín‐Guerrero, J. D., Micheli, A., & Soria‐Olivas, E. (2017) · CINECA IRIS Institutial research information system (University of Pisa) · Cited by 35
Neural algorithms and computing beyond Moore's law
Aimone, J. B. (2019) · Communications of the ACM · doi:10.1145/3231589 · Cited by 45
“Increasingly, algorithms like liquid state machines (LSMs) 25 have been introduced that abstractly emulate cortical dynamics loosely by balancing activity in neural circuits that exhibit chaotic (or near chaotic) activity.”
BackgroundGroßberger, L., Battaglia, F. P., & Vinck, M. A. (2018) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1006283 · Cited by 43
Performance Improvement of Chaotic Baseband Wireless Communication Using Echo State Network
Ren, H., Yin, H., Bai, C., & Yao, J. (2020) · IEEE Transactions on Communications · doi:10.1109/tcomm.2020.3007757 · Cited by 35
“Reservoir computing is a machine-learning approach that has been successfully applied for different tasks, [44], [45] including time series prediction.”
BackgroundMethodLearning unseen coexisting attractors
Gauthier, D. J., Fischer, I., & Röhm, A. (2022) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0116784 · Cited by 38
Reservoir computing using networks of memristors: effects of topology and heterogeneity
Mallinson, J. B., Heywood, Z. E., Daniels, R. K., Arnold, M. D., Bones, P. J., & Brown, S. A. (2023) · Nanoscale · doi:10.1039/d2nr07275k · Cited by 38
Neocortical layer 4 as a pluripotent function linearizer
Favorov, O. V., & Kurşun, O. (2011) · Journal of Neurophysiology · doi:10.1152/jn.00708.2010 · Cited by 36
“2009) strongly supports the proposals (Poggio 1990; Pouget and Sejnowski 1997; Maass et al. 2002; DiCarlo and Cox 2007) that such a strategy might also be used by the neocortex in the experiencedriven development of its remarkable functional properties (Sur and Rubenstein 2005).”
MethodResultSpace-Time Algebra: A Model for Neocortical Computation
Smith, J. E. (2018) · Proceedings - International Symposium on Computer Architecture/Proceedings · doi:10.1109/isca.2018.00033 · Cited by 43
“[33][44] are based on the same principles as TNNs: temporal encoding and spiking neuron models.”
MethodAnalysis of the ReSuMe Learning Process For Spiking Neural Networks
Ponulak, F. (2008) · International Journal of Applied Mathematics and Computer Science · doi:10.2478/v10006-008-0011-1 · Cited by 35
“In our analysis we consider the ReSuMe learning algorithm implemented in the LSM (Liquid State Machine) network architecture (Maass, Natschlaeger and Markram, 2002; Natschlaeger, Maass and Markram, 2002).”
MethodThe Capacity of Quantum Neural Networks
Wright, L. G., & McMahon, P. L. (2019) · arXiv · doi:10.48550/arxiv.1908.01364 · Cited by 35
Huang, W., Zhang, H., Tang, J., Lin, Z., Guo, T., Zhou, Y., Jiang, S., Hang, P., Jiao, M., Zhu, C., et al. (2024) · ACS Photonics · doi:10.1021/acsphotonics.4c00379 · Cited by 35
Kikumoto, A., Bhandari, A., Shibata, K., & Badre, D. (2024) · Nature Communications · doi:10.1038/s41467-024-52777-6 · Cited by 38
“For example, recurrent neural networks that use chaotic dynamics or reservoir computing can achieve exceptionally high context sensitivity, yet they are temporally unstable 15 – 19 .”
BackgroundData-driven model discovery with Kolmogorov-Arnold networks
Panahi, S., Moradi, M., Bollt, E. M., & Lai, Y. (2025) · Physical Review Research · doi:10.1103/physrevresearch.7.023037 · Cited by 42
Method for Training a Spiking Neuron to Associate Input-Output Spike Trains
Mohemmed, A. W., Schliebs, S., Matsuda, S., & Kasabov, N. K. (2011) · IFIP International Federation for Information Processing/IFIP · doi:10.1007/978-3-642-23957-1_25 · Cited by 34
“Although the method can only train a single layer of the SNN, it was argued that in combination with the LSM approach [9], the method allows the processing of complex non-linearly separable classification problems.”
MethodLinking non-binned spike train kernels to several existing spike train metrics
Schrauwen, B., & Campenhout, J. V. (2006) · Neurocomputing · doi:10.1016/j.neucom.2006.11.017 · Cited by 39
“Recent metrics [2, 5, 7, 9] take these biological factors into account which results in better performance.”
BackgroundMethodResultInfluentialShayani, H., Bentley, P. J., & Tyrrell, A. M. (2008) · NASA/ESA Conference on Adaptive Hardware and Systems (AHS) · doi:10.1109/ahs.2008.13 · Cited by 34
“Recently, with independent works of Buonomano [10], Maass (Liquid State Machine - LSM [11]), Jaeger (Echo State Network - ESN [12]) and Steil (Back-Propagation Decorrelation - BPDC [13]) a new technique, collectively known as Reservoir Computing (RC) [14], emerged, which is claimed to be capable of processing analogue…”
MethodLi, X., Wang, W. J., Xue, F., & Song, Y. (2017) · Physica A Statistical Mechanics and its Applications · doi:10.1016/j.physa.2017.08.053 · Cited by 35
All Optical Integrate and Fire Neuromorphic Node Based on Single Section Quantum Dot Laser
Sarantoglou, G., Skontranis, M., & Mesaritakis, C. (2019) · IEEE Journal of Selected Topics in Quantum Electronics · doi:10.1109/jstqe.2019.2945549 · Cited by 36
“brid information representation, provides significant advantages such as resilience to noise accumulation [14], spatio-temporal encoding of information [15], [16] and it can provide significant performance enhancement in specific class of machine learning”
BackgroundVariational Online Learning of Neural Dynamics
Zhao, Y., & Park, I. M. (2020) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2020.00071 · Cited by 43
“Chaotic Dynamics Chaotic dynamics (or edge-of-chaos) have been postulated to support asynchronous states in the cortex and neural computation over time by generating rich temporal patterns (Maass et al., 2002; Laje and Buonomano, 2013).”
BackgroundBenchmarking the performance of neuromorphic and spiking neural network simulators
Kulkarni, S., Parsa, M., Mitchell, J. P., & Schuman, C. D. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.03.028 · Cited by 34
Nonlinear input transformations are ubiquitous in quantum reservoir computing
Govia, L. C. G., Ribeill, G. J., Rowlands, G. E., & Ohki, T. (2022) · Neuromorphic Computing and Engineering · doi:10.1088/2634-4386/ac4fcd · Cited by 45
Kumar, S. D. V., Lo, M. Y. K., Arumugam, T., & Karuppanan, S. (2021) · Materials · doi:10.3390/ma14206135 · Cited by 34
“Echo state networks (ESNs) Similar to FFNN but utilizes random connections within the nodes [46].”
MethodAdaptive time-delayed photonic reservoir computing based on Kalman-filter training
Jin, J., Jiang, N., Zhang, Y., Feng, W., Zhao, A., Liu, S., Peng, J., Qiu, K. L., & Zhang, Q. (2022) · Optics Express · doi:10.1364/oe.454852 · Cited by 38
“The RC has been extensively applied in many areas, such as financial time-series prediction, nonlinear channel equalization [3,4], human action recognition [5], and reconstruction of complex dynamic system [6], etc.”
MethodAutomotive Radar Processing With Spiking Neural Networks: Concepts and Challenges
Vogginger, B., Kreutz, F., López-Randulfe, J., Liu, C., Dietrich, R., Gonzalez, H. A., Scholz, D., Reeb, N., Auge, D., Hille, J., et al. (2022) · Frontiers in Neuroscience · doi:10.3389/fnins.2022.851774 · Cited by 38
“Common general approaches for that are the neural engineering framework (Eliasmith and Anderson, 2003) or liquid state machines (Maass et al., 2002).”
BackgroundAudio Classification with Skyrmion Reservoirs
Msiska, R., Love, J., Mulkers, J., Leliaert, J., & Everschor‐Sitte, K. (2023) · Advanced Intelligent Systems · doi:10.1002/aisy.202200388 · Cited by 37
“Initial experiments in RC revolved around the usage of artificial neural network reservoirs, such as echo state networks [1] and liquid state machines [2], where the reservoir consists of a large number of individual interconnected nodes or neurons.”
A novel approach to minimal reservoir computing
Ma, H., Prosperino, D., & Rath, C. W. (2023) · Scientific Reports · doi:10.1038/s41598-023-39886-w · Cited by 41
“A reservoir computer (RC) 5,23,24 is an artificial recurrent neural network (RNN) that relies on a static network called reservoir .”
BackgroundInput Dependent Cell Assembly Dynamics in a Model of the Striatal Medium Spiny Neuron Network
Ponzi, A. P. D., & Wickens, J. R. (2012) · Frontiers in Systems Neuroscience · doi:10.3389/fnsys.2012.00006 · Cited by 36
“Moreover the dual encoding of stimulus identity and elapsed time in a single spiking network (Buonomano and Merzenich, 1995; Maass et al., 2002; Jaeger and Hass, 2004; Karmarkar and Buonomano, 2007; Buonomano and Maass, 2009) may be relevant in the temporal credit assignment tasks which are thought to deeply involve t…”
BackgroundMeng, Y. M., Jin, Y., & Yin, J. (2011) · IEEE Transactions on Neural Networks · doi:10.1109/tnn.2011.2171044 · Cited by 37
Comparing evolutionary methods for reservoir computing pre-training
Ferreira, A. A., & Ludermir, T. B. (2011) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2011.6033233 · Cited by 40
“A new RNN paradigm, suggested by Verstraeten, Schrauwen, D'Haen e Stroobandt [11], named Reservoir Computing (RC), proposes the union of LSMs and ESNs in a single research field.”
BackgroundDasgupta, S., Wörgötter, F., & Manoonpong, P. (2014) · Frontiers in Neural Circuits · doi:10.3389/fncir.2014.00126 · Cited by 35
“Inspired by the reservoir computing framework (Maass et al., 2002; Jaeger and Haas, 2004), here we use a chaotic random recurrent neural network (RNN) (Sussillo and Abbott, 2009; Rajan et al., 2010) as the adaptive critic (cortico-striatal circuitry and the DA system) connected to a feed-forward…”
MethodBig Data Analytics for Program Popularity Prediction in Broadcast TV Industries
Zhu, C., Cheng, G., & Wang, K. (2017) · IEEE Access · doi:10.1109/access.2017.2767104 · Cited by 37
Learning to Recognize Actions From Limited Training Examples Using a Recurrent Spiking Neural Model
Panda, P., & Srinivasa, N. (2018) · Frontiers in Neuroscience · doi:10.3389/fnins.2018.00126 · Cited by 42
“In order to quantify the effectiveness of the D/A approach, we compared the performance of our D/A based reservoir with that of a vanilla Liquid State Machine (LSM) (Maass et al., 2002; Maass, 2011).”
ResultInput Prediction and Autonomous Movement Analysis in Recurrent Circuits of Spikitfg Neurons
Legenstein, R. A., Markram, Η., & Maass, W. (2003) · Reviews in the Neurosciences · doi:10.1515/revneuro.2003.14.1-2.5 · Cited by 36
Self-citationTechnical report: supervised training of convolutional spiking neural networks with PyTorchPreprint
Zimmer, R., Pellegrini, T., Singh, S. F., & Masquelier, T. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1911.10124 · Cited by 34
A Comparative Study of Reservoir Computing for Temporal Signal ProcessingPreprint
Goudarzi, A., Banda, P., Lakin, M. R., Teuscher, C., & Stefanović, D. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1401.2224 · Cited by 41
“[14] proved that given the two properties of separation and approximation, a reservoir system is capable of approximating any time series.”
BackgroundReinforcement learning of linking and tracing contours in recurrent neural networks
Brosch, T., Neumann, H., & Roelfsema, P. R. (2015) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1004489 · Cited by 33
“RELEARNN also differs from reservoir computing methods, such as echo state networks [116, 117], liquid state networks [120] and backpropagation-decorrelation learning [121].”
BackgroundWhy the Realism Debate Matters for Science Policy: The Case of the Human Brain Project
Shaw, J. (2018) · Spontaneous Generations A Journal for the History and Philosophy of Science · doi:10.4245/sponge.v9i1.27760 · Cited by 33
Čindrak, S., Donvil, B., Lüdge, K., & Jaurigue, L. (2024) · Physical Review Research · doi:10.1103/physrevresearch.6.013051 · Cited by 46
“The scheme is based on the fading memory property usually assumed for reservoir computers [10, 11], and the memory-nonlinearity trade-off which is known to occur in reservoir computing [27–29].”
BackgroundWang, H., & Yan, X. (2014) · Neurocomputing · doi:10.1016/j.neucom.2014.05.024 · Cited by 33
An adaptive decoder from spike trains to micro-stimulation using kernel least-mean-squares (KLMS)
Li, L., Park, I. M., Seth, S., Choi, J. S., Francis, J. T., Sanchez, J. C., & Príncipe, J. C. (2011) · Machine learning for signal processing ... · doi:10.1109/mlsp.2011.6064603 · Cited by 34
Panda, P., & Roy, K. (2017) · Frontiers in Neuroscience · doi:10.3389/fnins.2017.00693 · Cited by 33
“, with class labels) to extract the information from the reservoir (Maass et al., 2002; Sussillo and Abbott, 2009).”
BackgroundA Computational Perspective of the Role of the Thalamus in Cognition
Dehghani, N., & Wimmer, R. D. (2019) · Neural Computation · doi:10.1162/neco_a_01197 · Cited by 35
A Survey of Neuromorphic Computing Based on Spiking Neural Networks
Zhang, M., Gu, Z., & Pan, G. (2018) · Chinese Journal of Electronics · doi:10.1049/cje.2018.05.006 · Cited by 36
Li, N., Tuo, J., Wang, Y., & Wang, M. (2019) · Neurocomputing · doi:10.1016/j.neucom.2019.10.003 · Cited by 33
Neuromorphic Implementation of a Recurrent Neural Network for EMG Classification
Ma, Y., Donati, E., Chen, B., Ren, P., Zheng, N., & Indiveri, G. (2020) · IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS) · doi:10.1109/aicas48895.2020.9073810 · Cited by 33
“The SNN setup on DYNAP-SE is a LSM.”
BackgroundMethodInfluentialLai, Y.-C. (2021) · Chaos An Interdisciplinary Journal of Nonlinear Science · doi:10.1063/5.0062042 · Cited by 33
Second-order information bottleneck based spiking neural networks for sEMG recognition
Zhang, A., Niu, Y., Gao, Y., Wu, J., & Gao, Z. (2021) · Information Sciences · doi:10.1016/j.ins.2021.11.065 · Cited by 34
Uthamacumaran, A., Elouatik, S., Abdouh, M., Berteau‐Rainville, M., Gao, Z., & Arena, G. O. (2022) · Applied Intelligence · doi:10.1007/s10489-022-03203-1 · Cited by 35
“Recurrent Neural Networks such as Reservoir Computing and liquid neural networks are available as tools for causal pattern discovery in complex time-series datasets which could be exploited in time-series EVs spectroscopic measurements and longitudinal blood monitoring [48–51].”
Liquid State Machine on SpiNNaker for Spatio-Temporal Classification Tasks
Patiño-Saucedo, A., Rostro‐González, H., Serrano‐Gotarredona, T., & Linares-Barranco, B. (2022) · Frontiers in Neuroscience · doi:10.3389/fnins.2022.819063 · Cited by 36
Integrated photonic reservoir computing with an all-optical readout
Ma, C., Kerrebrouck, J. V., Deng, H., Sackesyn, S., Gooskens, E., Bai, B., Dambre, J., & Bienstman, P. (2023) · Optics Express · doi:10.1364/oe.502354 · Cited by 33
“In this context, photonic reservoir computing (RC) [1, 2] has emerged as a promising computational architecture for optical information processing, offering next-generation high bandwidth, low latency, and low power consumption capabilities.”
BackgroundLearning shapes neural geometry in the primate prefrontal cortex
Wójcik, M. J., Stroud, J. P., Wasmuht, D., Kusunoki, M., Kadohisa, M., Buckley, M. J., Costa, R. P., Myers, N. E., Hunt, L. T., Duncan, J., et al. (2026) · Nature Neuroscience · doi:10.1038/s41593-026-02333-w · Cited by 33
Markovian Bias of Neural-based Architectures With Feedback Connections
Peter Tiňo, B., Hammer, B., & Mikael Bodén, A. (2007) · Studies in computational intelligence · doi:10.1007/978-3-540-73954-8_5 · Cited by 33
“It has been shown in [29] that LSMs are approximation complete for operators with fading memory under mild conditions on the models.”
BackgroundDynamic stability of sequential stimulus representations in adapting neuronal networks
Duarte, R., & Morrison, A. (2014) · Frontiers in Computational Neuroscience · doi:10.3389/fncom.2014.00124 · Cited by 35
“This response matrix is then sampled at time point t0 = 200 ms, resulting in the N-dimensional vector xu(t0) which contains all the neurons’ responses to input pattern u, or the circuit state (Maass et al., 2002).”
BackgroundMesquita, D. P. P., Gomes, J. P. P., Rodrigues, L. R., Oliveira, S. A. F., & Galvão, R. K. H. (2017) · Applied Soft Computing · doi:10.1016/j.asoc.2017.08.007 · Cited by 34
Accelerating photonic computing by bandwidth enhancement of a time‐delay reservoir
Estébanez, I., Schwind, J., Fischer, I., & Argyris, A. (2018) · Nanophotonics · doi:10.1515/nanoph-2020-0184 · Cited by 37
Behavior switching using reservoir computing for a soft robotic arm
Li, T., Nakajima, K., Cianchetti, M., Laschi, C., & Pfeifer, R. (2012) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2012.6225366 · Cited by 33
“Reservoir computing [16], [17], [18] is a new way to construct and train the RNNs.”
Learn more by training less: systematicity in sentence processing by recurrent networks
Frank, S. L. (2006) · Connection Science · doi:10.1080/09540090600768336 · Cited by 33
“Indeed, independently from each other, Jaeger (2003) and Maass et al. (2002) proposed a new type of recurrent neural network based on the idea that the recurrent, dynamical part of the network does not require training.”
BackgroundDistributions of covariances as a window into the operational regime of neuronal networksPreprint
Dahmen, D., Diesmann, M., & Helias, M. (2016) · JuSER (Forschungszentrum Jülich) · doi:10.48550/arxiv.1605.04153 · Cited by 34
Donati, G. P., Argyris, A., Mancinelli, M., Mirasso, C. R., & Pavesi, L. (2024) · Optics Express · doi:10.1364/oe.514617 · Cited by 32
“Reservoir Computing (RC) is a popular hardware-friendly paradigm that makes the training of the neural network time-and cost-efficient [5,6].”
BackgroundThe Kernel Adaptive Autoregressive-Moving-Average Algorithm
Li, K., & Príncipe, J. C. (2015) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2015.2418323 · Cited by 32
“This reservoir computing (RC) framework is further identified as echo state networks [21] for continuous valued inputs and liquid state machines (LSMs) [22] for spike train inputs.”
BackgroundCorrelated Fluctuations in Strongly Coupled Binary Networks Beyond Equilibrium
Dahmen, D., Bos, H., & Helias, M. (2016) · Physical Review X · doi:10.1103/physrevx.6.031024 · Cited by 42
“This insight has consequences for the use of random networks with fixed weights in the framework of reservoir computing [77,78], where highest computational performance [79] is achieved at the edge of chaos (for a review, see Ref.”
BackgroundSSO-LSM: A Sparse and Self-Organizing architecture for Liquid State Machine based neural processors
Jin, Y., Liu, Y., & Li, P. (2016) · International Symposium on Nanoscale Architectures · doi:10.1145/2950067.2950100 · Cited by 40
“The liquid state machine (LSM) is a specific form of reservoir computing, which is based on spiking neurons [9].”
BackgroundZhang, X., Foderaro, G., Henriquez, C. S., & Ferrari, S. (2016) · International Journal of Neural Systems · doi:10.1142/s0129065717500150 · Cited by 32
Assessing the Health of LiFePO4 Traction Batteries through Monotonic Echo State Networks
Sánchez, L., Anseán, D., Otero, J., & Couso, I. (2017) · Sensors · doi:10.3390/s18010009 · Cited by 33
Neural dynamics of emotion and cognition: From trajectories to underlying neural geometry
Pessoa, L. (2019) · Neural Networks · doi:10.1016/j.neunet.2019.08.007 · Cited by 37
Distributed Kerr Non-linearity in a Coherent All-Optical Fiber-Ring Reservoir Computer
Pauwels, J., Verschaffelt, G., Massar, S., & der Sande, G. V. (2019) · Frontiers in Physics · doi:10.3389/fphy.2019.00138 · Cited by 34
“Reservoir Computing The framework of reservoir computing allows to exploit the transient non-linear dynamics of a dynamical system to perform useful computation [1, 2].”
BackgroundBig Data Processing Technologies in Distributed Information Systems
Shakhovska, N., Boyko, N. I., Zasoba, Y., & Beňová, E. (2019) · Procedia Computer Science · doi:10.1016/j.procs.2019.11.047 · Cited by 35
Periodic structure of memory function in spintronics reservoir with feedback current
Yamaguchi, T., Akashi, N., Tsunegi, S., Kubota, H., Nakajima, K., & Taniguchi, T. (2020) · Physical Review Research · doi:10.1103/physrevresearch.2.023389 · Cited by 34
Büchel, J., Zendrikov, D., Solinas, S. M. G., Indiveri, G., & Muir, D. R. (2021) · Scientific Reports · doi:10.1038/s41598-021-02779-x · Cited by 38
Forecasting High-Frequency Spatio-Temporal Wind Power with Dimensionally Reduced Echo State Networks
Huang, H., Castruccio, S., & Genton, M. G. (2022) · Journal of the Royal Statistical Society Series C (Applied Statistics) · doi:10.1111/rssc.12540 · Cited by 32
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.