Biologically-inspired neuronal adaptation improves learning in neural networks

References

• Mnih V, Kavukcuoglu K, Silver D, et al. Human-level control through deep reinforcement learning. Nature. 2015;518(7540):529–9.
   PubMed Web of Science ®Google Scholar
• Silver D, Huang A, Maddison CJ, et al. Mastering the Game of Go with deep neural networks and tree search. Nature. 2016;529(7587):484–489.
   PubMed Web of Science ®Google Scholar
• Rumelhart DE, Hinton GE, Williams RJ. Learning internal representations by error propagation. La Jolla, CA: California Univ San Diego La Jolla Inst for Cognitive Science; 1985.
   Google Scholar
• Crick F. The recent excitement about neural networks. Nature. 1989;337(6203):129–132.
   PubMed Web of Science ®Google Scholar
• Lillicrap TP, Santoro A, Marris L, et al. Backpropagation and the brain. Nat Rev Neurosci. 2020;21(6):335–346.
   PubMed Web of Science ®Google Scholar
• Bengio Y. How auto-encoders could provide credit assignment in deep networks via target propagation. arXiv preprint arXiv. 2014:1407.7906.
   Google Scholar
• Hinton GE, McClelland J. In Proceedings of the 1987 International Conference on Neural information processing systems; 1987 Jan 1. Learning representations by recirculation; p. 358–366.
   Google Scholar
• Lecun Y. Modeles connexionnistes de l’apprentissage (connectionist learning models) [PhD thesis]; 1987.
   Google Scholar
• Lillicrap TP, Cownden D, Tweed DB, et al. Random synaptic feedback weights support error backpropagation for deep learning. Nat Commun. 2016;7(1):1–10.
   Google Scholar
• Movellan JR. Contrastive Hebbian learning in the continuous Hopfield model. In: Morgan Kaufmann, editor. Connectionist models. San Francisco, CA: Elsevier; 1991 Jan 1. p. 10–17.
   Google Scholar
• O’Reilly RC. Biologically plausible error-driven learning using local activation differences: the generalized recirculation algorithm. Neural Comput. 1996;8(5):895–938.
   Web of Science ®Google Scholar
• Scellier B, Bengio Y. Equilibrium propagation: bridging the gap between energy-based models and backpropagation. Front Comput Neurosci. 2017;11:24.
   PubMed Web of Science ®Google Scholar
• Almeida LB. A learning rule for asynchronous perceptrons with feedback in a combinatorial environment. In: Caudil M and Butler C, Editors. Proceedings of the IEEE First International Conference on Neural Networks San Diego, CA; 1987. p. 609–618.
   Google Scholar
• Baldi P, Pineda F. Contrastive learning and neural oscillations. Neural Comput. 1991;3(4):526–545.
   PubMed Web of Science ®Google Scholar
• Pineda FJ. Generalization of back-propagation to recurrent neural networks. Phys Rev Lett. 1987;59(19):2229.
   PubMed Web of Science ®Google Scholar
• Ernoult M, Grollier J, Querlioz D, Bengio Y, Scellier B. Updates of equilibrium prop match gradients of backprop through time in an RNN with static input. Adv Neural Inf Process Syst. 2019;32.
   Google Scholar
• Laborieux A, Ernoult M, Scellier B, et al. Scaling equilibrium propagation to deep convnets by drastically reducing its gradient estimator bias. Front Neurosci. 2021;15:129.
   Web of Science ®Google Scholar
• Scellier B, Bengio Y. Equivalence of equilibrium propagation and recurrent backpropagation. Neural Comput. 2019;31(2):312–329.
   PubMed Web of Science ®Google Scholar
• Luczak A, McNaughton BL, Kubo Y. Neurons learn by predicting future activity. Nature Mach Intell. 2022;4(1):1–11.
   Google Scholar
• Benda J. Neural adaptation. Curr Biol. 2021;31(3):R110–R116.
   PubMed Web of Science ®Google Scholar
• Whitmire CJ, Stanley GB. Rapid sensory adaptation redux: a circuit perspective. Neuron. 2016;92(2):298–315.
   PubMed Web of Science ®Google Scholar
• Hertã¤g L, Durstewitz D, Brunel N. Analytical approximations of the firing rate of an adaptive exponential integrate-and-fire neuron in the presence of synaptic noise. Front Comput Neurosci. 2014;8:116.
   PubMed Web of Science ®Google Scholar
• Jolivet R, Rauch A, Lüscher HR, Gerstner W. Integrate-and-fire models with adaptation are good enough. Adv Neural Inf Process Syst. 2005;18:595–602.
   Google Scholar
• Reutimann J, Yakovlev V, Fusi S, Senn W. Climbing neuronal activity as an event-based cortical representation of time. J Neurosci. 2004;24(13):3295–3303.
   PubMed Web of Science ®Google Scholar
• Stemmler M, Koch C. How voltage-dependent conductances can adapt to maximize the information encoded by neuronal firing rate. Nat Neurosci. 1999;2(6):521–527.
   PubMed Web of Science ®Google Scholar
• Fontaine B, Peña JL, Brette R. Spike-threshold adaptation predicted by membrane potential dynamics in vivo. PLoS Comput Biol. 2014;10(4):e1003560.
   PubMed Web of Science ®Google Scholar
• Carandini M, Ferster D. Membrane potential and firing rate in cat primary visual cortex. J Neurosci. 2000;20(1):470–484.
   PubMed Web of Science ®Google Scholar
• Granit R, Kernell D, Shortess G. Quantitative aspects of repetitive firing of mammalian motoneurones, caused by injected currents. J Physiol. 1963;168(4):911.
   PubMed Web of Science ®Google Scholar
• Treves A. Learning to predict through adaptation. Neuroinformatics. 2004;2(3):361–365.
   PubMed Web of Science ®Google Scholar
• Treves A. Computational constraints that may have favoured the lamination of sensory cortex. J Comput Neurosci. 2003;14(3):271–282.
   PubMed Web of Science ®Google Scholar
• Treves A. Computational constraints between retrieving the past and predicting the future, and the CA3‐CA1 differentiation. Hippocampus. 2004;14(5):539–556.
   PubMed Web of Science ®Google Scholar
• Treves A. Frontal latching networks: a possible neural basis for infinite recursion. Cogn Neuropsychol. 2005;22(3–4):276–291.
   PubMed Web of Science ®Google Scholar
• Luczak A, Kubo Y. Predictive neuronal adaptation as a basis for consciousness. Front Syst Neurosci. 2021;15. DOI:10.3389/fnsys.2021.767461
   PubMed Web of Science ®Google Scholar
• LeCun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition. Proc IEEE. 1998;86(11):2278–2324.
   Web of Science ®Google Scholar
• Krizhevsky A, Hinton G. Learning multiple layers of features from tiny images; 2009.
   Google Scholar
• Li F-F. CS231n: convolutional neural networks for visual recognition; 2021 [cited 2021 October 1]. Available from: https://cs231n.github.io/neural-networks-3/#annealing-the-learning-rate
   Google Scholar
• Sun J, Niu Z, Innanen KA, Li J, Trad D. A deep learning perspective of the forward and inverse problems in exploration geophysics. CSEG Geoconvention; 2019.
   Google Scholar
• Li Y, Wei C, Ma T. Towards explaining the regularization effect of initial large learning rate in training neural networks. Adv Neural Inf Process Syst. 2019;33:32.
   Google Scholar
• Duchi J, Hazan E, Singer Y. Adaptive subgradient methods for online learning and stochastic optimization. J Mach Learn Res. 2011;12(7):2121–2159.
   Google Scholar
• Lillicrap TP, Cownden D, Tweed DB, Akerman CJ. Random synaptic feedback weights support error backpropagation for deep learning. Nature communications. 2016;7(1): 1–10.
   Google Scholar
• Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R. Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res. 2014;15(1):1929–1958.
   Google Scholar
• Xie X, Seung HS. Equivalence of backpropagation and contrastive Hebbian learning in a layered network. Neural Comput. 2003;15(2):441–454.
   PubMed Web of Science ®Google Scholar
• Luczak A, Hackett TA, Kajikawa Y, et al. Multivariate receptive field mapping in marmoset auditory cortex. J Neurosci Methods. 2004;136(1):77–85.
   PubMed Web of Science ®Google Scholar
• Luczak A, Narayanan NS. Spectral representation—analyzing single-unit activity in extracellularly recorded neuronal data without spike sorting. J Neurosci Methods. 2005;144(1):53–61.
   PubMed Web of Science ®Google Scholar
• Ponjavic-Conte KD, et al. Neural correlates of auditory distraction revealed in theta-band EEG. Neuroreport: 2012;23(4):240–245.
   PubMed Web of Science ®Google Scholar
• Ryait H, Bermudez-Contreras E, Harvey M, et al. Data-driven analyses of motor impairments in animal models of neurological disorders. PLoS Biol. 2019;17(11):e3000516.
   PubMed Web of Science ®Google Scholar
• Schjetnan AGP, Luczak A. Recording large-scale neuronal ensembles with silicon probes in the anesthetized rat. J Vis Exp. 2011;56:e3282.
   Google Scholar
• Chalmers E, Luczak A. Reinforcement learning with brain-inspired modulation can improve adaptation to environmental changes. arXiv preprint arXiv. 2022;2205.09729.
   Google Scholar
• Kubo Y, Chalmers E, Luczak A. Combining backpropagation with equilibrium propagation to improve an actor-critic reinforcement learning framework. Front Comput Neurosci. 2022;16:980613.
   PubMed Web of Science ®Google Scholar