Learning with “Relevance”: Using a Third Factor to Stabilize Hebbian Learning

Authors:
Bernd Porr;Florentin Wörgötter
Affiliations:
Department of Electronics and Electrical Engineering, University of Glasgow, Glasgow, GT12 8LT, Scotland B.Porr@elec.gla.ac.uk;Bernstein Centre for Computational Neuroscience, University of Göttingen, 37073 Göttingen, Germany worgott@bccn-goettingen.de
Venue:
Neural Computation
Year:
2007

Citing 15
Cited 8

A drive-reinforcement model of single neuron function: An alternative to the Hebbian neuronal model

AIP Conference Proceedings 151 on Neural Networks for Computing
Differential Hebbian learning

AIP Conference Proceedings 151 on Neural Networks for Computing
Self-Organization in a Perceptual Network

Computer
Technical Note: \cal Q-Learning

Machine Learning
A bottom up approach towards the acquisition and expression of sequential representions applied to a behaving real-world device: Distributed Adaptive Control III

Neural Networks - Special issue on neural control and robotics: biology and technology
Feedback control systems (4th ed.)

Feedback control systems (4th ed.)
Dopamine-dependent plasticity of corticostriatal synapses

Neural Networks - Computational models of neuromodulation
Learning to Predict by the Methods of Temporal Differences

Machine Learning
Isotropic sequence order learning

Neural Computation
ISO learning approximates a solution to the inverse-controller problem in an unsupervised behavioral paradigm

Neural Computation
How the shape of pre- and postsynaptic signals can influence STDP: a biophysical model

Neural Computation
Temporal Sequence Learning, Prediction, and Control: A Review of Different Models and Their Relation to Biological Mechanisms

Neural Computation
Temporal Difference Model Reproduces Anticipatory Neural Activity

Neural Computation
Strongly improved stability and faster convergence of temporal sequence learning by using input correlations only

Neural Computation
Strongly improved stability and faster convergence of temporal sequence learning by using input correlations only

Neural Computation

Second Order Conditioning in the Sub-cortical Nuclei of the Limbic System

SAB '08 Proceedings of the 10th international conference on Simulation of Adaptive Behavior: From Animals to Animats
A spiking neural network model of an actor-critic learning agent

Neural Computation
On the asymptotic equivalence between differential Hebbian and temporal difference learning

Neural Computation
Biomimetic approach to tacit learning based on compound control

IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
Learning and Reversal Learning in the Subcortical Limbic System: A Computational Model

Adaptive Behavior - Animals, Animats, Software Agents, Robots, Adaptive Systems
From modulated Hebbian plasticity to simple behavior learning through noise and weight saturation

Neural Networks
Solving the distal reward problem with rare correlations

Neural Computation
Novel method for using Q-learning in small microcontrollers

Proceedings of the 51st ACM Southeast Conference

Quantified Score

Hi-index	0.00

Visualization

Abstract

It is a well-known fact that Hebbian learning is inherently unstable because of its self-amplifying terms: the more a synapse grows, the stronger the postsynaptic activity, and therefore the faster the synaptic growth. This unwanted weight growth is driven by the autocorrelation term of Hebbian learning where the same synapse drives its own growth. On the other hand, the cross-correlation term performs actual learning where different inputs are correlated with each other. Consequently, we would like to minimize the autocorrelation and maximize the cross-correlation. Here we show that we can achieve this with a third factor that switches on learning when the autocorrelation is minimal or zero and the cross-correlation is maximal. The biological counterpart of such a third factor is a neuromodulator that switches on learning at a certain moment in time. We show in a behavioral experiment that our three-factor learning clearly outperforms classical Hebbian learning.