1
|
Maslennikov OV, Gao C, Nekorkin VI. Internal dynamics of recurrent neural networks trained to generate complex spatiotemporal patterns. CHAOS (WOODBURY, N.Y.) 2023; 33:093125. [PMID: 37722673 DOI: 10.1063/5.0166359] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Received: 07/04/2023] [Accepted: 08/28/2023] [Indexed: 09/20/2023]
Abstract
How complex patterns generated by neural systems are represented in individual neuronal activity is an essential problem in computational neuroscience as well as machine learning communities. Here, based on recurrent neural networks in the form of feedback reservoir computers, we show microscopic features resulting in generating spatiotemporal patterns including multicluster and chimera states. We show the effect of individual neural trajectories as well as whole-network activity distributions on exhibiting particular regimes. In addition, we address the question how trained output weights contribute to the autonomous multidimensional dynamics.
Collapse
Affiliation(s)
- Oleg V Maslennikov
- Federal Research Center A.V. Gaponov-Grekhov Institute of Applied Physics of the Russian Academy of Sciences, Nizhny Novgorod, Russia
| | - Chao Gao
- School of Artificial Intelligence, Optics and Electronics, Northwestern Polytechnical University, Xian, China
| | - Vladimir I Nekorkin
- Federal Research Center A.V. Gaponov-Grekhov Institute of Applied Physics of the Russian Academy of Sciences, Nizhny Novgorod, Russia
| |
Collapse
|
2
|
Cimeša L, Ciric L, Ostojic S. Geometry of population activity in spiking networks with low-rank structure. PLoS Comput Biol 2023; 19:e1011315. [PMID: 37549194 PMCID: PMC10461857 DOI: 10.1371/journal.pcbi.1011315] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/25/2022] [Revised: 08/28/2023] [Accepted: 06/27/2023] [Indexed: 08/09/2023] Open
Abstract
Recurrent network models are instrumental in investigating how behaviorally-relevant computations emerge from collective neural dynamics. A recently developed class of models based on low-rank connectivity provides an analytically tractable framework for understanding of how connectivity structure determines the geometry of low-dimensional dynamics and the ensuing computations. Such models however lack some fundamental biological constraints, and in particular represent individual neurons in terms of abstract units that communicate through continuous firing rates rather than discrete action potentials. Here we examine how far the theoretical insights obtained from low-rank rate networks transfer to more biologically plausible networks of spiking neurons. Adding a low-rank structure on top of random excitatory-inhibitory connectivity, we systematically compare the geometry of activity in networks of integrate-and-fire neurons to rate networks with statistically equivalent low-rank connectivity. We show that the mean-field predictions of rate networks allow us to identify low-dimensional dynamics at constant population-average activity in spiking networks, as well as novel non-linear regimes of activity such as out-of-phase oscillations and slow manifolds. We finally exploit these results to directly build spiking networks that perform nonlinear computations.
Collapse
Affiliation(s)
- Ljubica Cimeša
- Laboratoire de Neurosciences Cognitives Computationnelles, Département d’Études Cognitives, École Normale Supérieure, INSERM U960, PSL University, Paris, France
| | - Lazar Ciric
- Laboratoire de Neurosciences Cognitives Computationnelles, Département d’Études Cognitives, École Normale Supérieure, INSERM U960, PSL University, Paris, France
| | - Srdjan Ostojic
- Laboratoire de Neurosciences Cognitives Computationnelles, Département d’Études Cognitives, École Normale Supérieure, INSERM U960, PSL University, Paris, France
| |
Collapse
|
3
|
Hutt A, Rich S, Valiante TA, Lefebvre J. Intrinsic neural diversity quenches the dynamic volatility of neural networks. Proc Natl Acad Sci U S A 2023; 120:e2218841120. [PMID: 37399421 PMCID: PMC10334753 DOI: 10.1073/pnas.2218841120] [Citation(s) in RCA: 7] [Impact Index Per Article: 7.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/03/2022] [Accepted: 05/19/2023] [Indexed: 07/05/2023] Open
Abstract
Heterogeneity is the norm in biology. The brain is no different: Neuronal cell types are myriad, reflected through their cellular morphology, type, excitability, connectivity motifs, and ion channel distributions. While this biophysical diversity enriches neural systems' dynamical repertoire, it remains challenging to reconcile with the robustness and persistence of brain function over time (resilience). To better understand the relationship between excitability heterogeneity (variability in excitability within a population of neurons) and resilience, we analyzed both analytically and numerically a nonlinear sparse neural network with balanced excitatory and inhibitory connections evolving over long time scales. Homogeneous networks demonstrated increases in excitability, and strong firing rate correlations-signs of instability-in response to a slowly varying modulatory fluctuation. Excitability heterogeneity tuned network stability in a context-dependent way by restraining responses to modulatory challenges and limiting firing rate correlations, while enriching dynamics during states of low modulatory drive. Excitability heterogeneity was found to implement a homeostatic control mechanism enhancing network resilience to changes in population size, connection probability, strength and variability of synaptic weights, by quenching the volatility (i.e., its susceptibility to critical transitions) of its dynamics. Together, these results highlight the fundamental role played by cell-to-cell heterogeneity in the robustness of brain function in the face of change.
Collapse
Affiliation(s)
- Axel Hutt
- Université de Strasbourg, CNRS, Inria, ICube, MLMS, MIMESIS, StrasbourgF-67000, France
| | - Scott Rich
- Krembil Brain Institute, Division of Clinical and Computational Neuroscience, University Health Network, Toronto, ONM5T 0S8, Canada
| | - Taufik A. Valiante
- Krembil Brain Institute, Division of Clinical and Computational Neuroscience, University Health Network, Toronto, ONM5T 0S8, Canada
- Department of Electrical and Computer Engineering, University of Toronto, Toronto, ONM5S 3G8, Canada
- Institute of Biomedical Engineering, University of Toronto, Toronto, ONM5S 3G9, Canada
- Institute of Medical Sciences, University of Toronto, Toronto, ONM5S 1A8, Canada
- Division of Neurosurgery, Department of Surgery, University of Toronto, Toronto, ONM5G 2C4, Canada
- Center for Advancing Neurotechnological Innovation to Application, University of Toronto, Toronto, ONM5G 2A2, Canada
- Max Planck-University of Toronto Center for Neural Science and Technology, University of Toronto, Toronto, ONM5S 3G8, Canada
| | - Jérémie Lefebvre
- Krembil Brain Institute, Division of Clinical and Computational Neuroscience, University Health Network, Toronto, ONM5T 0S8, Canada
- Department of Biology, University of Ottawa, Ottawa, ONK1N 6N5, Canada
- Department of Mathematics, University of Toronto, Toronto, ONM5S 2E4, Canada
| |
Collapse
|
4
|
Arthur BJ, Kim CM, Chen S, Preibisch S, Darshan R. A scalable implementation of the recursive least-squares algorithm for training spiking neural networks. Front Neuroinform 2023; 17:1099510. [PMID: 37441157 PMCID: PMC10333503 DOI: 10.3389/fninf.2023.1099510] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/15/2022] [Accepted: 06/05/2023] [Indexed: 07/15/2023] Open
Abstract
Training spiking recurrent neural networks on neuronal recordings or behavioral tasks has become a popular way to study computations performed by the nervous system. As the size and complexity of neural recordings increase, there is a need for efficient algorithms that can train models in a short period of time using minimal resources. We present optimized CPU and GPU implementations of the recursive least-squares algorithm in spiking neural networks. The GPU implementation can train networks of one million neurons, with 100 million plastic synapses and a billion static synapses, about 1,000 times faster than an unoptimized reference CPU implementation. We demonstrate the code's utility by training a network, in less than an hour, to reproduce the activity of > 66, 000 recorded neurons of a mouse performing a decision-making task. The fast implementation enables a more interactive in-silico study of the dynamics and connectivity underlying multi-area computations. It also admits the possibility to train models as in-vivo experiments are being conducted, thus closing the loop between modeling and experiments.
Collapse
Affiliation(s)
- Benjamin J. Arthur
- Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, United States
| | - Christopher M. Kim
- Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, United States
- Laboratory of Biological Modeling, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD, United States
| | - Susu Chen
- Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, United States
| | - Stephan Preibisch
- Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, United States
| | - Ran Darshan
- Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, United States
| |
Collapse
|
5
|
Kim CM, Finkelstein A, Chow CC, Svoboda K, Darshan R. Distributing task-related neural activity across a cortical network through task-independent connections. Nat Commun 2023; 14:2851. [PMID: 37202424 DOI: 10.1038/s41467-023-38529-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/18/2022] [Accepted: 05/05/2023] [Indexed: 05/20/2023] Open
Abstract
Task-related neural activity is widespread across populations of neurons during goal-directed behaviors. However, little is known about the synaptic reorganization and circuit mechanisms that lead to broad activity changes. Here we trained a subset of neurons in a spiking network with strong synaptic interactions to reproduce the activity of neurons in the motor cortex during a decision-making task. Task-related activity, resembling the neural data, emerged across the network, even in the untrained neurons. Analysis of trained networks showed that strong untrained synapses, which were independent of the task and determined the dynamical state of the network, mediated the spread of task-related activity. Optogenetic perturbations suggest that the motor cortex is strongly-coupled, supporting the applicability of the mechanism to cortical networks. Our results reveal a cortical mechanism that facilitates distributed representations of task-variables by spreading the activity from a subset of plastic neurons to the entire network through task-independent strong synapses.
Collapse
Affiliation(s)
- Christopher M Kim
- Laboratory of Biological Modeling, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD, USA.
- Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA.
| | - Arseny Finkelstein
- Department of Physiology and Pharmacology, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
- Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel
| | - Carson C Chow
- Laboratory of Biological Modeling, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD, USA
| | - Karel Svoboda
- Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA
- Allen Institute for Neural Dynamics, Seattle, WA, USA
| | - Ran Darshan
- Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, VA, USA.
| |
Collapse
|
6
|
DePasquale B, Sussillo D, Abbott LF, Churchland MM. The centrality of population-level factors to network computation is demonstrated by a versatile approach for training spiking networks. Neuron 2023; 111:631-649.e10. [PMID: 36630961 PMCID: PMC10118067 DOI: 10.1016/j.neuron.2022.12.007] [Citation(s) in RCA: 13] [Impact Index Per Article: 13.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/21/2020] [Revised: 06/17/2022] [Accepted: 12/05/2022] [Indexed: 01/12/2023]
Abstract
Neural activity is often described in terms of population-level factors extracted from the responses of many neurons. Factors provide a lower-dimensional description with the aim of shedding light on network computations. Yet, mechanistically, computations are performed not by continuously valued factors but by interactions among neurons that spike discretely and variably. Models provide a means of bridging these levels of description. We developed a general method for training model networks of spiking neurons by leveraging factors extracted from either data or firing-rate-based networks. In addition to providing a useful model-building framework, this formalism illustrates how reliable and continuously valued factors can arise from seemingly stochastic spiking. Our framework establishes procedures for embedding this property in network models with different levels of realism. The relationship between spikes and factors in such networks provides a foundation for interpreting (and subtly redefining) commonly used quantities such as firing rates.
Collapse
Affiliation(s)
- Brian DePasquale
- Princeton Neuroscience Institute, Princeton University, Princeton NJ, USA; Department of Neuroscience, Columbia University, New York, NY, USA; Center for Theoretical Neuroscience, Columbia University, New York, NY, USA.
| | - David Sussillo
- Department of Electrical Engineering, Stanford University, Stanford, CA, USA; Wu Tsai Neurosciences Institute, Stanford University, Stanford, CA, USA
| | - L F Abbott
- Department of Neuroscience, Columbia University, New York, NY, USA; Center for Theoretical Neuroscience, Columbia University, New York, NY, USA; Zuckerman Mind Brain Behavior Institute, Columbia University, New York, NY, USA; Department of Physiology and Cellular Biophysics, Columbia University, New York, NY, USA; Kavli Institute for Brain Science, Columbia University, New York, NY, USA
| | - Mark M Churchland
- Department of Neuroscience, Columbia University, New York, NY, USA; Zuckerman Mind Brain Behavior Institute, Columbia University, New York, NY, USA; Kavli Institute for Brain Science, Columbia University, New York, NY, USA; Grossman Center for the Statistics of Mind, Columbia University, New York, NY, USA
| |
Collapse
|
7
|
Input correlations impede suppression of chaos and learning in balanced firing-rate networks. PLoS Comput Biol 2022; 18:e1010590. [DOI: 10.1371/journal.pcbi.1010590] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/21/2022] [Revised: 12/15/2022] [Accepted: 09/20/2022] [Indexed: 12/12/2022] Open
Abstract
Neural circuits exhibit complex activity patterns, both spontaneously and evoked by external stimuli. Information encoding and learning in neural circuits depend on how well time-varying stimuli can control spontaneous network activity. We show that in firing-rate networks in the balanced state, external control of recurrent dynamics, i.e., the suppression of internally-generated chaotic variability, strongly depends on correlations in the input. A distinctive feature of balanced networks is that, because common external input is dynamically canceled by recurrent feedback, it is far more difficult to suppress chaos with common input into each neuron than through independent input. To study this phenomenon, we develop a non-stationary dynamic mean-field theory for driven networks. The theory explains how the activity statistics and the largest Lyapunov exponent depend on the frequency and amplitude of the input, recurrent coupling strength, and network size, for both common and independent input. We further show that uncorrelated inputs facilitate learning in balanced networks.
Collapse
|
8
|
Pruning recurrent neural networks replicates adolescent changes in working memory and reinforcement learning. Proc Natl Acad Sci U S A 2022; 119:e2121331119. [PMID: 35622896 DOI: 10.1073/pnas.2121331119] [Citation(s) in RCA: 11] [Impact Index Per Article: 5.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/27/2022] Open
Abstract
SignificanceAdolescence is a period during which there are important changes in behavior and the structure of the brain. In this manuscript, we use theoretical modeling to show how improvements in working memory and reinforcement learning that occur during adolescence can be explained by the reduction in synaptic connectivity in prefrontal cortex that occurs during a similar period. We train recurrent neural networks to solve working memory and reinforcement learning tasks and show that when we prune connectivity in these networks, they perform the tasks better. The improvement in task performance, however, can come at the cost of flexibility as the pruned networks are not able to learn some new tasks as well.
Collapse
|
9
|
Intrinsic bursts facilitate learning of Lévy flight movements in recurrent neural network models. Sci Rep 2022; 12:4951. [PMID: 35322813 PMCID: PMC8943163 DOI: 10.1038/s41598-022-08953-z] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/08/2021] [Accepted: 03/09/2022] [Indexed: 11/24/2022] Open
Abstract
Isolated spikes and bursts of spikes are thought to provide the two major modes of information coding by neurons. Bursts are known to be crucial for fundamental processes between neuron pairs, such as neuronal communications and synaptic plasticity. Neuronal bursting also has implications in neurodegenerative diseases and mental disorders. Despite these findings on the roles of bursts, whether and how bursts have an advantage over isolated spikes in the network-level computation remains elusive. Here, we demonstrate in a computational model that not isolated spikes, but intrinsic bursts can greatly facilitate learning of Lévy flight random walk trajectories by synchronizing burst onsets across a neural population. Lévy flight is a hallmark of optimal search strategies and appears in cognitive behaviors such as saccadic eye movements and memory retrieval. Our results suggest that bursting is crucial for sequence learning by recurrent neural networks when sequences comprise long-tailed distributed discrete jumps.
Collapse
|
10
|
Toker D, Pappas I, Lendner JD, Frohlich J, Mateos DM, Muthukumaraswamy S, Carhart-Harris R, Paff M, Vespa PM, Monti MM, Sommer FT, Knight RT, D'Esposito M. Consciousness is supported by near-critical slow cortical electrodynamics. Proc Natl Acad Sci U S A 2022; 119:e2024455119. [PMID: 35145021 PMCID: PMC8851554 DOI: 10.1073/pnas.2024455119] [Citation(s) in RCA: 46] [Impact Index Per Article: 23.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/02/2020] [Accepted: 12/20/2021] [Indexed: 12/21/2022] Open
Abstract
Mounting evidence suggests that during conscious states, the electrodynamics of the cortex are poised near a critical point or phase transition and that this near-critical behavior supports the vast flow of information through cortical networks during conscious states. Here, we empirically identify a mathematically specific critical point near which waking cortical oscillatory dynamics operate, which is known as the edge-of-chaos critical point, or the boundary between stability and chaos. We do so by applying the recently developed modified 0-1 chaos test to electrocorticography (ECoG) and magnetoencephalography (MEG) recordings from the cortices of humans and macaques across normal waking, generalized seizure, anesthesia, and psychedelic states. Our evidence suggests that cortical information processing is disrupted during unconscious states because of a transition of low-frequency cortical electric oscillations away from this critical point; conversely, we show that psychedelics may increase the information richness of cortical activity by tuning low-frequency cortical oscillations closer to this critical point. Finally, we analyze clinical electroencephalography (EEG) recordings from patients with disorders of consciousness (DOC) and show that assessing the proximity of slow cortical oscillatory electrodynamics to the edge-of-chaos critical point may be useful as an index of consciousness in the clinical setting.
Collapse
Affiliation(s)
- Daniel Toker
- Department of Psychology, University of California, Los Angeles, CA 90095;
| | - Ioannis Pappas
- Helen Wills Neuroscience Institute, University of California, Berkeley, CA 94704
- Department of Psychology, University of California, Berkeley, CA 94704
- Laboratory of Neuro Imaging, Stevens Institute for Neuroimaging and Informatics, Keck School of Medicine, University of Southern California, Los Angeles, CA 90033
| | - Janna D Lendner
- Helen Wills Neuroscience Institute, University of California, Berkeley, CA 94704
- Department of Anesthesiology and Intensive Care, University Medical Center, 72076 Tübingen, Germany
| | - Joel Frohlich
- Department of Psychology, University of California, Los Angeles, CA 90095
| | - Diego M Mateos
- Consejo Nacional de Investigaciones Científicas y Técnicas de Argentina, C1425 Buenos Aires, Argentina
- Facultad de Ciencia y Tecnología, Universidad Autónoma de Entre Ríos, E3202 Paraná, Entre Ríos, Argentina
- Grupo de Análisis de Neuroimágenes, Instituo de Matemática Aplicada del Litoral, S3000 Santa Fe, Argentina
| | - Suresh Muthukumaraswamy
- School of Pharmacy, Faculty of Medical and Health Sciences, The University of Auckland, 1010 Auckland, New Zealand
| | - Robin Carhart-Harris
- Neuropsychopharmacology Unit, Centre for Psychiatry, Imperial College London, London SW7 2AZ, United Kingdom
- Centre for Psychedelic Research, Department of Psychiatry, Imperial College London, London SW7 2AZ, United Kingdom
| | - Michelle Paff
- Department of Neurological Surgery, University of California, Irvine, CA 92697
| | - Paul M Vespa
- Brain Injury Research Center, Department of Neurosurgery, University of California, Los Angeles, CA 90095
| | - Martin M Monti
- Department of Psychology, University of California, Los Angeles, CA 90095
- Brain Injury Research Center, Department of Neurosurgery, University of California, Los Angeles, CA 90095
| | - Friedrich T Sommer
- Helen Wills Neuroscience Institute, University of California, Berkeley, CA 94704
- Redwood Center for Theoretical Neuroscience, University of California, Berkeley, CA 94704
| | - Robert T Knight
- Helen Wills Neuroscience Institute, University of California, Berkeley, CA 94704
- Department of Psychology, University of California, Berkeley, CA 94704
| | - Mark D'Esposito
- Helen Wills Neuroscience Institute, University of California, Berkeley, CA 94704
- Department of Psychology, University of California, Berkeley, CA 94704
| |
Collapse
|
11
|
Dasbach S, Tetzlaff T, Diesmann M, Senk J. Dynamical Characteristics of Recurrent Neuronal Networks Are Robust Against Low Synaptic Weight Resolution. Front Neurosci 2021; 15:757790. [PMID: 35002599 PMCID: PMC8740282 DOI: 10.3389/fnins.2021.757790] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/12/2021] [Accepted: 11/03/2021] [Indexed: 11/13/2022] Open
Abstract
The representation of the natural-density, heterogeneous connectivity of neuronal network models at relevant spatial scales remains a challenge for Computational Neuroscience and Neuromorphic Computing. In particular, the memory demands imposed by the vast number of synapses in brain-scale network simulations constitute a major obstacle. Limiting the number resolution of synaptic weights appears to be a natural strategy to reduce memory and compute load. In this study, we investigate the effects of a limited synaptic-weight resolution on the dynamics of recurrent spiking neuronal networks resembling local cortical circuits and develop strategies for minimizing deviations from the dynamics of networks with high-resolution synaptic weights. We mimic the effect of a limited synaptic weight resolution by replacing normally distributed synaptic weights with weights drawn from a discrete distribution, and compare the resulting statistics characterizing firing rates, spike-train irregularity, and correlation coefficients with the reference solution. We show that a naive discretization of synaptic weights generally leads to a distortion of the spike-train statistics. If the weights are discretized such that the mean and the variance of the total synaptic input currents are preserved, the firing statistics remain unaffected for the types of networks considered in this study. For networks with sufficiently heterogeneous in-degrees, the firing statistics can be preserved even if all synaptic weights are replaced by the mean of the weight distribution. We conclude that even for simple networks with non-plastic neurons and synapses, a discretization of synaptic weights can lead to substantial deviations in the firing statistics unless the discretization is performed with care and guided by a rigorous validation process. For the network model used in this study, the synaptic weights can be replaced by low-resolution weights without affecting its macroscopic dynamical characteristics, thereby saving substantial amounts of memory.
Collapse
Affiliation(s)
- Stefan Dasbach
- Institute of Neuroscience and Medicine (INM-6) and Institute for Advanced Simulation (IAS-6) and JARA Institute Brain Structure-Function Relationships (INM-10), Jülich Research Centre, Jülich, Germany
| | - Tom Tetzlaff
- Institute of Neuroscience and Medicine (INM-6) and Institute for Advanced Simulation (IAS-6) and JARA Institute Brain Structure-Function Relationships (INM-10), Jülich Research Centre, Jülich, Germany
| | - Markus Diesmann
- Institute of Neuroscience and Medicine (INM-6) and Institute for Advanced Simulation (IAS-6) and JARA Institute Brain Structure-Function Relationships (INM-10), Jülich Research Centre, Jülich, Germany
- Department of Physics, Faculty 1, RWTH Aachen University, Aachen, Germany
- Department of Psychiatry, Psychotherapy, and Psychosomatics, Medical School, RWTH Aachen University, Aachen, Germany
| | - Johanna Senk
- Institute of Neuroscience and Medicine (INM-6) and Institute for Advanced Simulation (IAS-6) and JARA Institute Brain Structure-Function Relationships (INM-10), Jülich Research Centre, Jülich, Germany
| |
Collapse
|
12
|
Bird AD, Jedlicka P, Cuntz H. Dendritic normalisation improves learning in sparsely connected artificial neural networks. PLoS Comput Biol 2021; 17:e1009202. [PMID: 34370727 PMCID: PMC8407571 DOI: 10.1371/journal.pcbi.1009202] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/10/2020] [Revised: 08/31/2021] [Accepted: 06/19/2021] [Indexed: 11/25/2022] Open
Abstract
Artificial neural networks, taking inspiration from biological neurons, have become an invaluable tool for machine learning applications. Recent studies have developed techniques to effectively tune the connectivity of sparsely-connected artificial neural networks, which have the potential to be more computationally efficient than their fully-connected counterparts and more closely resemble the architectures of biological systems. We here present a normalisation, based on the biophysical behaviour of neuronal dendrites receiving distributed synaptic inputs, that divides the weight of an artificial neuron's afferent contacts by their number. We apply this dendritic normalisation to various sparsely-connected feedforward network architectures, as well as simple recurrent and self-organised networks with spatially extended units. The learning performance is significantly increased, providing an improvement over other widely-used normalisations in sparse networks. The results are two-fold, being both a practical advance in machine learning and an insight into how the structure of neuronal dendritic arbours may contribute to computation.
Collapse
Affiliation(s)
- Alex D. Bird
- Ernst Strüngmann Institute for Neuroscience (ESI) in co-operation with Max Planck Society, Frankfurt, Germany
- Frankfurt Institute for Advanced Studies (FIAS), Frankfurt, Germany
- ICAR3R-Interdisciplinary Centre for 3Rs in Animal Research, Faculty of Medicine, Justus Liebig University Giessen, Giessen, Germany
| | - Peter Jedlicka
- Frankfurt Institute for Advanced Studies (FIAS), Frankfurt, Germany
- ICAR3R-Interdisciplinary Centre for 3Rs in Animal Research, Faculty of Medicine, Justus Liebig University Giessen, Giessen, Germany
| | - Hermann Cuntz
- Ernst Strüngmann Institute for Neuroscience (ESI) in co-operation with Max Planck Society, Frankfurt, Germany
- Frankfurt Institute for Advanced Studies (FIAS), Frankfurt, Germany
| |
Collapse
|
13
|
Transfer-RLS method and transfer-FORCE learning for simple and fast training of reservoir computing models. Neural Netw 2021; 143:550-563. [PMID: 34304003 DOI: 10.1016/j.neunet.2021.06.031] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/10/2020] [Revised: 05/13/2021] [Accepted: 06/29/2021] [Indexed: 11/22/2022]
Abstract
Reservoir computing is a machine learning framework derived from a special type of recurrent neural network. Following recent advances in physical reservoir computing, some reservoir computing devices are thought to be promising as energy-efficient machine learning hardware for real-time information processing. To realize efficient online learning with low-power reservoir computing devices, it is beneficial to develop fast convergence learning methods with simpler operations. This study proposes a training method located in the middle between the recursive least squares (RLS) method and the least mean squares (LMS) method, which are standard online learning methods for reservoir computing models. The RLS method converges fast but requires updates of a huge matrix called a gain matrix, whereas the LMS method does not use a gain matrix but converges very slow. On the other hand, the proposed method called a transfer-RLS method does not require updates of the gain matrix in the main-training phase by updating that in advance (i.e., in a pre-training phase). As a result, the transfer-RLS method can work with simpler operations than the original RLS method without sacrificing much convergence speed. We numerically and analytically show that the transfer-RLS method converges much faster than the LMS method. Furthermore, we show that a modified version of the transfer-RLS method (called transfer-FORCE learning) can be applied to the first-order reduced and controlled error (FORCE) learning for a reservoir computing model with a closed-loop, which is challenging to train.
Collapse
|
14
|
Lin X, Zou X, Ji Z, Huang T, Wu S, Mi Y. A brain-inspired computational model for spatio-temporal information processing. Neural Netw 2021; 143:74-87. [PMID: 34091238 DOI: 10.1016/j.neunet.2021.05.015] [Citation(s) in RCA: 8] [Impact Index Per Article: 2.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/14/2020] [Revised: 03/31/2021] [Accepted: 05/12/2021] [Indexed: 11/28/2022]
Abstract
Spatio-temporal information processing is fundamental in both brain functions and AI applications. Current strategies for spatio-temporal pattern recognition usually involve explicit feature extraction followed by feature aggregation, which requires a large amount of labeled data. In the present study, motivated by the subcortical visual pathway and early stages of the auditory pathway for motion and sound processing, we propose a novel brain-inspired computational model for generic spatio-temporal pattern recognition. The model consists of two modules, a reservoir module and a decision-making module. The former projects complex spatio-temporal patterns into spatially separated neural representations via its recurrent dynamics, the latter reads out neural representations via integrating information over time, and the two modules are linked together using known examples. Using synthetic data, we demonstrate that the model can extract the frequency and order information of temporal inputs. We apply the model to reproduce the looming pattern discrimination behavior as observed in experiments successfully. Furthermore, we apply the model to the gait recognition task, and demonstrate that our model accomplishes the recognition in an event-based manner and outperforms deep learning counterparts when training data is limited.
Collapse
Affiliation(s)
- Xiaohan Lin
- School of Electronics Engineering and Computer Science, Peking University, No.5 Yiheyuan Road Haidian District, Beijing 100871, PR China.
| | - Xiaolong Zou
- School of Electronics Engineering and Computer Science, Peking University, No.5 Yiheyuan Road Haidian District, Beijing 100871, PR China; School of Psychological and Cognitive Sciences, IDG/McGovern Institute for Brain Research, PKU-Tsinghua Center for Life Sciences, Peking University, No.5 Yiheyuan Road Haidian District, Beijing 100871, PR China.
| | - Zilong Ji
- School of Electronics Engineering and Computer Science, Peking University, No.5 Yiheyuan Road Haidian District, Beijing 100871, PR China; School of Psychological and Cognitive Sciences, IDG/McGovern Institute for Brain Research, PKU-Tsinghua Center for Life Sciences, Peking University, No.5 Yiheyuan Road Haidian District, Beijing 100871, PR China.
| | - Tiejun Huang
- School of Electronics Engineering and Computer Science, Peking University, No.5 Yiheyuan Road Haidian District, Beijing 100871, PR China.
| | - Si Wu
- School of Electronics Engineering and Computer Science, Peking University, No.5 Yiheyuan Road Haidian District, Beijing 100871, PR China; School of Psychological and Cognitive Sciences, IDG/McGovern Institute for Brain Research, PKU-Tsinghua Center for Life Sciences, Peking University, No.5 Yiheyuan Road Haidian District, Beijing 100871, PR China.
| | - Yuanyuan Mi
- Center for Neurointelligence, School of Medicine, Chongqing University, No.174 Shazhengjie, Shapingba, Chongqing 400044, PR China; AI Research Center, Peng Cheng Laboratory, No.2, Xingke First Street, Nanshan District, Shenzhen 518005, PR China.
| |
Collapse
|
15
|
Kim CM, Chow CC. Training Spiking Neural Networks in the Strong Coupling Regime. Neural Comput 2021; 33:1199-1233. [PMID: 34496392 DOI: 10.1162/neco_a_01379] [Citation(s) in RCA: 3] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/24/2020] [Accepted: 11/23/2020] [Indexed: 11/04/2022]
Abstract
Recurrent neural networks trained to perform complex tasks can provide insight into the dynamic mechanism that underlies computations performed by cortical circuits. However, due to a large number of unconstrained synaptic connections, the recurrent connectivity that emerges from network training may not be biologically plausible. Therefore, it remains unknown if and how biological neural circuits implement dynamic mechanisms proposed by the models. To narrow this gap, we developed a training scheme that, in addition to achieving learning goals, respects the structural and dynamic properties of a standard cortical circuit model: strongly coupled excitatory-inhibitory spiking neural networks. By preserving the strong mean excitatory and inhibitory coupling of initial networks, we found that most of trained synapses obeyed Dale's law without additional constraints, exhibited large trial-to-trial spiking variability, and operated in inhibition-stabilized regime. We derived analytical estimates on how training and network parameters constrained the changes in mean synaptic strength during training. Our results demonstrate that training recurrent neural networks subject to strong coupling constraints can result in connectivity structure and dynamic regime relevant to cortical circuits.
Collapse
Affiliation(s)
- Christopher M Kim
- Laboratory of Biological Modeling, National Institute of Diabetes and Digestive and Kidney Diseases/National Institutes of Health, Bethesda, MD 20814, U.S.A.
| | - Carson C Chow
- Laboratory of Biological Modeling, National Institute of Diabetes and Digestive and Kidney Diseases/National Institutes of Health, Bethesda, MD 20814, U.S.A.
| |
Collapse
|
16
|
Muratore P, Capone C, Paolucci PS. Target spike patterns enable efficient and biologically plausible learning for complex temporal tasks. PLoS One 2021; 16:e0247014. [PMID: 33592040 PMCID: PMC7886200 DOI: 10.1371/journal.pone.0247014] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/25/2020] [Accepted: 01/31/2021] [Indexed: 11/28/2022] Open
Abstract
Recurrent spiking neural networks (RSNN) in the brain learn to perform a wide range of perceptual, cognitive and motor tasks very efficiently in terms of energy consumption and their training requires very few examples. This motivates the search for biologically inspired learning rules for RSNNs, aiming to improve our understanding of brain computation and the efficiency of artificial intelligence. Several spiking models and learning rules have been proposed, but it remains a challenge to design RSNNs whose learning relies on biologically plausible mechanisms and are capable of solving complex temporal tasks. In this paper, we derive a learning rule, local to the synapse, from a simple mathematical principle, the maximization of the likelihood for the network to solve a specific task. We propose a novel target-based learning scheme in which the learning rule derived from likelihood maximization is used to mimic a specific spatio-temporal spike pattern that encodes the solution to complex temporal tasks. This method makes the learning extremely rapid and precise, outperforming state of the art algorithms for RSNNs. While error-based approaches, (e.g. e-prop) trial after trial optimize the internal sequence of spikes in order to progressively minimize the MSE we assume that a signal randomly projected from an external origin (e.g. from other brain areas) directly defines the target sequence. This facilitates the learning procedure since the network is trained from the beginning to reproduce the desired internal sequence. We propose two versions of our learning rule: spike-dependent and voltage-dependent. We find that the latter provides remarkable benefits in terms of learning speed and robustness to noise. We demonstrate the capacity of our model to tackle several problems like learning multidimensional trajectories and solving the classical temporal XOR benchmark. Finally, we show that an online approximation of the gradient ascent, in addition to guaranteeing complete locality in time and space, allows learning after very few presentations of the target output. Our model can be applied to different types of biological neurons. The analytically derived plasticity learning rule is specific to each neuron model and can produce a theoretical prediction for experimental validation.
Collapse
Affiliation(s)
- Paolo Muratore
- SISSA—International School for Advanced Studies, Trieste, Italy
- * E-mail:
| | | | | |
Collapse
|
17
|
Berberian N, Ross M, Chartier S. Embodied working memory during ongoing input streams. PLoS One 2021; 16:e0244822. [PMID: 33400724 PMCID: PMC7785253 DOI: 10.1371/journal.pone.0244822] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/18/2020] [Accepted: 12/16/2020] [Indexed: 11/18/2022] Open
Abstract
Sensory stimuli endow animals with the ability to generate an internal representation. This representation can be maintained for a certain duration in the absence of previously elicited inputs. The reliance on an internal representation rather than purely on the basis of external stimuli is a hallmark feature of higher-order functions such as working memory. Patterns of neural activity produced in response to sensory inputs can continue long after the disappearance of previous inputs. Experimental and theoretical studies have largely invested in understanding how animals faithfully maintain sensory representations during ongoing reverberations of neural activity. However, these studies have focused on preassigned protocols of stimulus presentation, leaving out by default the possibility of exploring how the content of working memory interacts with ongoing input streams. Here, we study working memory using a network of spiking neurons with dynamic synapses subject to short-term and long-term synaptic plasticity. The formal model is embodied in a physical robot as a companion approach under which neuronal activity is directly linked to motor output. The artificial agent is used as a methodological tool for studying the formation of working memory capacity. To this end, we devise a keyboard listening framework to delineate the context under which working memory content is (1) refined, (2) overwritten or (3) resisted by ongoing new input streams. Ultimately, this study takes a neurorobotic perspective to resurface the long-standing implication of working memory in flexible cognition.
Collapse
Affiliation(s)
- Nareg Berberian
- Laboratory for Computational Neurodynamics and Cognition, School of Psychology, University of Ottawa, Ottawa, Ontario, Canada
| | - Matt Ross
- Laboratory for Computational Neurodynamics and Cognition, School of Psychology, University of Ottawa, Ottawa, Ontario, Canada
| | - Sylvain Chartier
- Laboratory for Computational Neurodynamics and Cognition, School of Psychology, University of Ottawa, Ottawa, Ontario, Canada
| |
Collapse
|
18
|
Ingrosso A. Optimal learning with excitatory and inhibitory synapses. PLoS Comput Biol 2020; 16:e1008536. [PMID: 33370266 PMCID: PMC7793294 DOI: 10.1371/journal.pcbi.1008536] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/27/2020] [Revised: 01/08/2021] [Accepted: 11/13/2020] [Indexed: 11/22/2022] Open
Abstract
Characterizing the relation between weight structure and input/output statistics is fundamental for understanding the computational capabilities of neural circuits. In this work, I study the problem of storing associations between analog signals in the presence of correlations, using methods from statistical mechanics. I characterize the typical learning performance in terms of the power spectrum of random input and output processes. I show that optimal synaptic weight configurations reach a capacity of 0.5 for any fraction of excitatory to inhibitory weights and have a peculiar synaptic distribution with a finite fraction of silent synapses. I further provide a link between typical learning performance and principal components analysis in single cases. These results may shed light on the synaptic profile of brain circuits, such as cerebellar structures, that are thought to engage in processing time-dependent signals and performing on-line prediction.
Collapse
Affiliation(s)
- Alessandro Ingrosso
- Zuckerman Mind, Brain, Behavior Institute, Columbia University, New York, New York, United States of America
| |
Collapse
|
19
|
Akbarzadeh-Sherbaf K, Safari S, Vahabie AH. A digital hardware implementation of spiking neural networks with binary FORCE training. Neurocomputing 2020. [DOI: 10.1016/j.neucom.2020.05.044] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/15/2023]
|
20
|
Baker C, Zhu V, Rosenbaum R. Nonlinear stimulus representations in neural circuits with approximate excitatory-inhibitory balance. PLoS Comput Biol 2020; 16:e1008192. [PMID: 32946433 PMCID: PMC7526938 DOI: 10.1371/journal.pcbi.1008192] [Citation(s) in RCA: 10] [Impact Index Per Article: 2.5] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/21/2020] [Revised: 09/30/2020] [Accepted: 07/24/2020] [Indexed: 12/02/2022] Open
Abstract
Balanced excitation and inhibition is widely observed in cortex. How does this balance shape neural computations and stimulus representations? This question is often studied using computational models of neuronal networks in a dynamically balanced state. But balanced network models predict a linear relationship between stimuli and population responses. So how do cortical circuits implement nonlinear representations and computations? We show that every balanced network architecture admits stimuli that break the balanced state and these breaks in balance push the network into a "semi-balanced state" characterized by excess inhibition to some neurons, but an absence of excess excitation. The semi-balanced state produces nonlinear stimulus representations and nonlinear computations, is unavoidable in networks driven by multiple stimuli, is consistent with cortical recordings, and has a direct mathematical relationship to artificial neural networks.
Collapse
Affiliation(s)
- Cody Baker
- Department of Applied and Computational Mathematics and Statistics, University of Notre Dame, Notre Dame, IN, USA
| | - Vicky Zhu
- Department of Applied and Computational Mathematics and Statistics, University of Notre Dame, Notre Dame, IN, USA
| | - Robert Rosenbaum
- Department of Applied and Computational Mathematics and Statistics, University of Notre Dame, Notre Dame, IN, USA
- Interdisciplinary Center for Network Science and Applications, University of Notre Dame, Notre Dame, IN, USA
| |
Collapse
|
21
|
A solution to the learning dilemma for recurrent networks of spiking neurons. Nat Commun 2020; 11:3625. [PMID: 32681001 PMCID: PMC7367848 DOI: 10.1038/s41467-020-17236-y] [Citation(s) in RCA: 93] [Impact Index Per Article: 23.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/09/2019] [Accepted: 06/16/2020] [Indexed: 11/09/2022] Open
Abstract
Recurrently connected networks of spiking neurons underlie the astounding information processing capabilities of the brain. Yet in spite of extensive research, how they can learn through synaptic plasticity to carry out complex network computations remains unclear. We argue that two pieces of this puzzle were provided by experimental data from neuroscience. A mathematical result tells us how these pieces need to be combined to enable biologically plausible online network learning through gradient descent, in particular deep reinforcement learning. This learning method-called e-prop-approaches the performance of backpropagation through time (BPTT), the best-known method for training recurrent neural networks in machine learning. In addition, it suggests a method for powerful on-chip learning in energy-efficient spike-based hardware for artificial intelligence.
Collapse
|
22
|
Kim R, Li Y, Sejnowski TJ. Simple framework for constructing functional spiking recurrent neural networks. Proc Natl Acad Sci U S A 2019; 116:22811-22820. [PMID: 31636215 DOI: 10.1101/579706] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 05/25/2023] Open
Abstract
Cortical microcircuits exhibit complex recurrent architectures that possess dynamically rich properties. The neurons that make up these microcircuits communicate mainly via discrete spikes, and it is not clear how spikes give rise to dynamics that can be used to perform computationally challenging tasks. In contrast, continuous models of rate-coding neurons can be trained to perform complex tasks. Here, we present a simple framework to construct biologically realistic spiking recurrent neural networks (RNNs) capable of learning a wide range of tasks. Our framework involves training a continuous-variable rate RNN with important biophysical constraints and transferring the learned dynamics and constraints to a spiking RNN in a one-to-one manner. The proposed framework introduces only 1 additional parameter to establish the equivalence between rate and spiking RNN models. We also study other model parameters related to the rate and spiking networks to optimize the one-to-one mapping. By establishing a close relationship between rate and spiking models, we demonstrate that spiking RNNs could be constructed to achieve similar performance as their counterpart continuous rate networks.
Collapse
Affiliation(s)
- Robert Kim
- Computational Neurobiology Laboratory, Salk Institute for Biological Studies, La Jolla, CA 92037;
- Neurosciences Graduate Program, University of California San Diego, La Jolla, CA 92093
- Medical Scientist Training Program, University of California San Diego, La Jolla, CA 92093
| | - Yinghao Li
- Computational Neurobiology Laboratory, Salk Institute for Biological Studies, La Jolla, CA 92037
| | - Terrence J Sejnowski
- Computational Neurobiology Laboratory, Salk Institute for Biological Studies, La Jolla, CA 92037;
- Institute for Neural Computation, University of California San Diego, La Jolla, CA 92093
- Division of Biological Sciences, University of California San Diego, La Jolla, CA 92093
| |
Collapse
|
23
|
Kim R, Li Y, Sejnowski TJ. Simple framework for constructing functional spiking recurrent neural networks. Proc Natl Acad Sci U S A 2019; 116:22811-22820. [PMID: 31636215 PMCID: PMC6842655 DOI: 10.1073/pnas.1905926116] [Citation(s) in RCA: 23] [Impact Index Per Article: 4.6] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/18/2022] Open
Abstract
Cortical microcircuits exhibit complex recurrent architectures that possess dynamically rich properties. The neurons that make up these microcircuits communicate mainly via discrete spikes, and it is not clear how spikes give rise to dynamics that can be used to perform computationally challenging tasks. In contrast, continuous models of rate-coding neurons can be trained to perform complex tasks. Here, we present a simple framework to construct biologically realistic spiking recurrent neural networks (RNNs) capable of learning a wide range of tasks. Our framework involves training a continuous-variable rate RNN with important biophysical constraints and transferring the learned dynamics and constraints to a spiking RNN in a one-to-one manner. The proposed framework introduces only 1 additional parameter to establish the equivalence between rate and spiking RNN models. We also study other model parameters related to the rate and spiking networks to optimize the one-to-one mapping. By establishing a close relationship between rate and spiking models, we demonstrate that spiking RNNs could be constructed to achieve similar performance as their counterpart continuous rate networks.
Collapse
Affiliation(s)
- Robert Kim
- Computational Neurobiology Laboratory, Salk Institute for Biological Studies, La Jolla, CA 92037;
- Neurosciences Graduate Program, University of California San Diego, La Jolla, CA 92093
- Medical Scientist Training Program, University of California San Diego, La Jolla, CA 92093
| | - Yinghao Li
- Computational Neurobiology Laboratory, Salk Institute for Biological Studies, La Jolla, CA 92037
| | - Terrence J Sejnowski
- Computational Neurobiology Laboratory, Salk Institute for Biological Studies, La Jolla, CA 92037;
- Institute for Neural Computation, University of California San Diego, La Jolla, CA 92093
- Division of Biological Sciences, University of California San Diego, La Jolla, CA 92093
| |
Collapse
|
24
|
Maslennikov OV, Nekorkin VI. Collective dynamics of rate neurons for supervised learning in a reservoir computing system. CHAOS (WOODBURY, N.Y.) 2019; 29:103126. [PMID: 31675797 DOI: 10.1063/1.5119895] [Citation(s) in RCA: 7] [Impact Index Per Article: 1.4] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 07/14/2019] [Accepted: 09/26/2019] [Indexed: 06/10/2023]
Abstract
In this paper, we study collective dynamics of the network of rate neurons which constitute a central element of a reservoir computing system. The main objective of the paper is to identify the dynamic behaviors inside the reservoir underlying the performance of basic machine learning tasks, such as generating patterns with specified characteristics. We build a reservoir computing system which includes a reservoir-a network of interacting rate neurons-and an output element that generates a target signal. We study individual activities of interacting rate neurons, while implementing the task and analyze the impact of the dynamic parameter-a time constant-on the quality of implementation.
Collapse
Affiliation(s)
- Oleg V Maslennikov
- Institute of Applied Physics of the Russian Academy of Sciences, 46 Ulyanov Street, 603950 Nizhny Novgorod, Russia
| | - Vladimir I Nekorkin
- Institute of Applied Physics of the Russian Academy of Sciences, 46 Ulyanov Street, 603950 Nizhny Novgorod, Russia
| |
Collapse
|
25
|
|
26
|
Training dynamically balanced excitatory-inhibitory networks. PLoS One 2019; 14:e0220547. [PMID: 31393909 PMCID: PMC6687153 DOI: 10.1371/journal.pone.0220547] [Citation(s) in RCA: 23] [Impact Index Per Article: 4.6] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/10/2019] [Accepted: 07/19/2019] [Indexed: 12/02/2022] Open
Abstract
The construction of biologically plausible models of neural circuits is crucial for understanding the computational properties of the nervous system. Constructing functional networks composed of separate excitatory and inhibitory neurons obeying Dale’s law presents a number of challenges. We show how a target-based approach, when combined with a fast online constrained optimization technique, is capable of building functional models of rate and spiking recurrent neural networks in which excitation and inhibition are balanced. Balanced networks can be trained to produce complicated temporal patterns and to solve input-output tasks while retaining biologically desirable features such as Dale’s law and response variability.
Collapse
|
27
|
Mastrogiuseppe F, Ostojic S. A Geometrical Analysis of Global Stability in Trained Feedback Networks. Neural Comput 2019; 31:1139-1182. [DOI: 10.1162/neco_a_01187] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.2] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/04/2022]
Abstract
Recurrent neural networks have been extensively studied in the context of neuroscience and machine learning due to their ability to implement complex computations. While substantial progress in designing effective learning algorithms has been achieved, a full understanding of trained recurrent networks is still lacking. Specifically, the mechanisms that allow computations to emerge from the underlying recurrent dynamics are largely unknown. Here we focus on a simple yet underexplored computational setup: a feedback architecture trained to associate a stationary output to a stationary input. As a starting point, we derive an approximate analytical description of global dynamics in trained networks, which assumes uncorrelated connectivity weights in the feedback and in the random bulk. The resulting mean-field theory suggests that the task admits several classes of solutions, which imply different stability properties. Different classes are characterized in terms of the geometrical arrangement of the readout with respect to the input vectors, defined in the high-dimensional space spanned by the network population. We find that such an approximate theoretical approach can be used to understand how standard training techniques implement the input-output task in finite-size feedback networks. In particular, our simplified description captures the local and the global stability properties of the target solution, and thus predicts training performance.
Collapse
Affiliation(s)
- Francesca Mastrogiuseppe
- Laboratoire de Neurosciences Cognitives et Computationelles, INSERM U960, and Laboratoire de Physique Statistique, CNRS UMR 8550, Ecole Normale Supérieure–PSL Research University, Paris 75005, France
| | - Srdjan Ostojic
- Laboratoire de Neurosciences Cognitives et Computationelles, INSERM U960, Ecole Normale Supérieure–PSL Research University, Paris 75005, France
| |
Collapse
|