1
|
Kubo Y, Chalmers E, Luczak A. Biologically-inspired neuronal adaptation improves learning in neural networks. Commun Integr Biol 2023; 16:2163131. [PMID: 36685291 PMCID: PMC9851208 DOI: 10.1080/19420889.2022.2163131] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/19/2023] Open
Abstract
Since humans still outperform artificial neural networks on many tasks, drawing inspiration from the brain may help to improve current machine learning algorithms. Contrastive Hebbian learning (CHL) and equilibrium propagation (EP) are biologically plausible algorithms that update weights using only local information (without explicitly calculating gradients) and still achieve performance comparable to conventional backpropagation. In this study, we augmented CHL and EP with Adjusted Adaptation, inspired by the adaptation effect observed in neurons, in which a neuron's response to a given stimulus is adjusted after a short time. We add this adaptation feature to multilayer perceptrons and convolutional neural networks trained on MNIST and CIFAR-10. Surprisingly, adaptation improved the performance of these networks. We discuss the biological inspiration for this idea and investigate why Neuronal Adaptation could be an important brain mechanism to improve the stability and accuracy of learning.
Collapse
Affiliation(s)
- Yoshimasa Kubo
- Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, AB, Canada,CONTACT Yoshimasa Kubo
| | - Eric Chalmers
- Department of Mathematics & Computing, Mount Royal University, Calgary, AB, Canada
| | - Artur Luczak
- Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, AB, Canada,Artur Luczak Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, AB, Canada
| |
Collapse
|
2
|
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. [PMID: 36082305 PMCID: PMC9446087 DOI: 10.3389/fncom.2022.980613] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/28/2022] [Accepted: 08/05/2022] [Indexed: 01/09/2023] Open
Abstract
Backpropagation (BP) has been used to train neural networks for many years, allowing them to solve a wide variety of tasks like image classification, speech recognition, and reinforcement learning tasks. But the biological plausibility of BP as a mechanism of neural learning has been questioned. Equilibrium Propagation (EP) has been proposed as a more biologically plausible alternative and achieves comparable accuracy on the CIFAR-10 image classification task. This study proposes the first EP-based reinforcement learning architecture: an Actor-Critic architecture with the actor network trained by EP. We show that this model can solve the basic control tasks often used as benchmarks for BP-based models. Interestingly, our trained model demonstrates more consistent high-reward behavior than a comparable model trained exclusively by BP.
Collapse
Affiliation(s)
- Yoshimasa Kubo
- Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, AB, Canada
| | - Eric Chalmers
- Department of Mathematics and Computing, Mount Royal University, Calgary, AB, Canada
| | - Artur Luczak
- Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Lethbridge, AB, Canada
| |
Collapse
|
3
|
Laborieux A, Ernoult M, Scellier B, Bengio Y, Grollier J, Querlioz D. Scaling Equilibrium Propagation to Deep ConvNets by Drastically Reducing Its Gradient Estimator Bias. Front Neurosci 2021; 15:633674. [PMID: 33679315 PMCID: PMC7930909 DOI: 10.3389/fnins.2021.633674] [Citation(s) in RCA: 9] [Impact Index Per Article: 3.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/25/2020] [Accepted: 01/26/2021] [Indexed: 11/24/2022] Open
Abstract
Equilibrium Propagation is a biologically-inspired algorithm that trains convergent recurrent neural networks with a local learning rule. This approach constitutes a major lead to allow learning-capable neuromophic systems and comes with strong theoretical guarantees. Equilibrium propagation operates in two phases, during which the network is let to evolve freely and then "nudged" toward a target; the weights of the network are then updated based solely on the states of the neurons that they connect. The weight updates of Equilibrium Propagation have been shown mathematically to approach those provided by Backpropagation Through Time (BPTT), the mainstream approach to train recurrent neural networks, when nudging is performed with infinitely small strength. In practice, however, the standard implementation of Equilibrium Propagation does not scale to visual tasks harder than MNIST. In this work, we show that a bias in the gradient estimate of equilibrium propagation, inherent in the use of finite nudging, is responsible for this phenomenon and that canceling it allows training deep convolutional neural networks. We show that this bias can be greatly reduced by using symmetric nudging (a positive nudging and a negative one). We also generalize Equilibrium Propagation to the case of cross-entropy loss (by opposition to squared error). As a result of these advances, we are able to achieve a test error of 11.7% on CIFAR-10, which approaches the one achieved by BPTT and provides a major improvement with respect to the standard Equilibrium Propagation that gives 86% test error. We also apply these techniques to train an architecture with unidirectional forward and backward connections, yielding a 13.2% test error. These results highlight equilibrium propagation as a compelling biologically-plausible approach to compute error gradients in deep neuromorphic systems.
Collapse
Affiliation(s)
- Axel Laborieux
- Université Paris-Saclay, CNRS, Centre de Nanosciences et de Nanotechnologies, Palaiseau, France
| | - Maxence Ernoult
- Université Paris-Saclay, CNRS, Centre de Nanosciences et de Nanotechnologies, Palaiseau, France
- Unité Mixte de Physique, CNRS, Thales, Université Paris-Saclay, Palaiseau, France
- Mila, Université de Montréal, Montreal, QC, Canada
| | | | - Yoshua Bengio
- Mila, Université de Montréal, Montreal, QC, Canada
- Canadian Institute for Advanced Research, Toronto, ON, Canada
| | - Julie Grollier
- Unité Mixte de Physique, CNRS, Thales, Université Paris-Saclay, Palaiseau, France
| | - Damien Querlioz
- Université Paris-Saclay, CNRS, Centre de Nanosciences et de Nanotechnologies, Palaiseau, France
| |
Collapse
|
4
|
Kruijne W, Bohte SM, Roelfsema PR, Olivers CNL. Flexible Working Memory Through Selective Gating and Attentional Tagging. Neural Comput 2020; 33:1-40. [PMID: 33080159 DOI: 10.1162/neco_a_01339] [Citation(s) in RCA: 13] [Impact Index Per Article: 3.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/02/2023]
Abstract
Working memory is essential: it serves to guide intelligent behavior of humans and nonhuman primates when task-relevant stimuli are no longer present to the senses. Moreover, complex tasks often require that multiple working memory representations can be flexibly and independently maintained, prioritized, and updated according to changing task demands. Thus far, neural network models of working memory have been unable to offer an integrative account of how such control mechanisms can be acquired in a biologically plausible manner. Here, we present WorkMATe, a neural network architecture that models cognitive control over working memory content and learns the appropriate control operations needed to solve complex working memory tasks. Key components of the model include a gated memory circuit that is controlled by internal actions, encoding sensory information through untrained connections, and a neural circuit that matches sensory inputs to memory content. The network is trained by means of a biologically plausible reinforcement learning rule that relies on attentional feedback and reward prediction errors to guide synaptic updates. We demonstrate that the model successfully acquires policies to solve classical working memory tasks, such as delayed recognition and delayed pro-saccade/anti-saccade tasks. In addition, the model solves much more complex tasks, including the hierarchical 12-AX task or the ABAB ordered recognition task, both of which demand an agent to independently store and updated multiple items separately in memory. Furthermore, the control strategies that the model acquires for these tasks subsequently generalize to new task contexts with novel stimuli, thus bringing symbolic production rule qualities to a neural network architecture. As such, WorkMATe provides a new solution for the neural implementation of flexible memory control.
Collapse
Affiliation(s)
- Wouter Kruijne
- Faculty of Behavior and Movement Sciences, Vrije Universiteit Amsterdam, 1081 BT Amsterdam, Noord Holland, The Netherlands
| | - Sander M Bohte
- Machine Learning Group, Centrum voor Wiskunde & Informatica, 1098 XG Amsterdam, Noord Holland, The Netherlands; Swammerdam Institute of Life Sciences, University of Amsterdam, 1098 XH Amsterdam, Noord Holland, The Netherlands; and Department of Computer Science, Rijksuniversiteit Groningen, 9747 AG Groningen, The Netherlands
| | - Pieter R Roelfsema
- Department of Vision & Cognition, Netherlands Institute for Neuroscience, 1105BA Amsterdam, Noord Holland, The Netherlands; Department of Integrative Neurophysiology, Center for Neurogenomics and Cognitive Research, Vrije Universiteit Amsterdam, 1981 HV Amsterdam, Noord Holland, The Netherlands; and Department of Computer Science, Rijksuniversiteit Groningen, 9747 AG Groningen, The Netherlands
| | - Christian N L Olivers
- Faculty of Behavior and Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam, Noord Holland, The Netherlands, Department of Psychiatry, Academic Medical Center, Amsterdam, The Netherlands
| |
Collapse
|
5
|
Whittington JCR, Bogacz R. Theories of Error Back-Propagation in the Brain. Trends Cogn Sci 2019; 23:235-250. [PMID: 30704969 PMCID: PMC6382460 DOI: 10.1016/j.tics.2018.12.005] [Citation(s) in RCA: 136] [Impact Index Per Article: 27.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/07/2018] [Revised: 12/13/2018] [Accepted: 12/28/2018] [Indexed: 12/14/2022]
Abstract
This review article summarises recently proposed theories on how neural circuits in the brain could approximate the error back-propagation algorithm used by artificial neural networks. Computational models implementing these theories achieve learning as efficient as artificial neural networks, but they use simple synaptic plasticity rules based on activity of presynaptic and postsynaptic neurons. The models have similarities, such as including both feedforward and feedback connections, allowing information about error to propagate throughout the network. Furthermore, they incorporate experimental evidence on neural connectivity, responses, and plasticity. These models provide insights on how brain networks might be organised such that modification of synaptic weights on multiple levels of cortical hierarchy leads to improved performance on tasks.
Collapse
Affiliation(s)
- James C R Whittington
- MRC Brain Network Dynamics Unit, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford OX3 9DU, UK; Wellcome Centre for Integrative Neuroimaging, Centre for Functional Magnetic Resonance Imaging of the Brain, University of Oxford, Oxford OX3 9DU, UK
| | - Rafal Bogacz
- MRC Brain Network Dynamics Unit, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford OX3 9DU, UK.
| |
Collapse
|