Synchronization and redundancy: implications for robustness of neural learning and decision making

DOI10.1162/NECO_A_00183zbMATH Open1237.92010arXiv1312.1632OpenAlexW1999698770WikidataQ49084582 ScholiaQ49084582MaRDI QIDQ2887012FDOQ2887012

Authors: Jake Bouvrie, Jean-Jacques Slotine 

Publication date: 15 May 2012

Published in: Neural Computation (Search for Journal in Brave)

Abstract: To learn and reason in the presence of uncertainty, the brain must be capable of imposing some form of regularization. Here we suggest, through theoretical and computational arguments, that the combination of noise with synchronization provides a plausible mechanism for regularization in the nervous system. The functional role of regularization is considered in a general context in which coupled computational systems receive inputs corrupted by correlated noise. Noise on the inputs is shown to impose regularization, and when synchronization upstream induces time-varying correlations across noise variables, the degree of regularization can be calibrated over time. The proposed mechanism is explored first in the context of a simple associative learning problem, and then in the context of a hierarchical sensory coding task. The resulting qualitative behavior coincides with experimental data from visual cortex.

Full work available at URL: https://arxiv.org/abs/1312.1632

Recommendations

• scientific article; zbMATH DE number 7275134
• Oscillations and Spiking Pairs: Behavior of a Neuronal Model with STDP Learning
• Synchronization of heteroclinic circuits through learning in coupled neural networks
• When Response Variability Increases Neural Network Robustness to Synaptic Noise
• Statistical mechanics of reward-modulated learning in decision-making networks

Mathematics Subject Classification ID

Probabilistic models, generic numerical methods in probability and statistics (65C20) Neural biology (92C20) Memory and learning in psychology (91E40) Biological rhythms and synchronization (92B25)

Cites Work

• On early stopping in gradient descent learning
• A Contraction Theory Approach to Stochastic Incremental Stability
• Coordination of groups of mobile autonomous agents using nearest neighbor rules
• On contraction analysis for non-linear systems
• On partial contraction analysis for coupled nonlinear oscillators
• Symmetry and phaselocking in chains of weakly coupled oscillators
• Collective enhancement of precision in networks of coupled oscillators
• Stable concurrent synchronization in dynamic system networks
• Convergence in human decision-making dynamics
• The neuronal replicator hypothesis

Cited In (3)

• Title not available (Why is that?)
• Title not available (Why is that?)
• Accuracy and response-time distributions for decision-making: linear perfect integrators versus nonlinear attractor-based neural circuits

This page was built for publication: Synchronization and redundancy: implications for robustness of neural learning and decision making

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2887012)