Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On
Boltzmann Machines
L. Shao. (2012)cite arxiv:1210.8442Comment: Submitted to International Conference of Learning Representation (ICLR) 2013.
Abstract
One conjecture in both deep learning and classical connectionist viewpoint is
that the biological brain implements certain kinds of deep networks as its
back-end. However, to our knowledge, a detailed correspondence has not yet been
set up, which is important if we want to bridge between neuroscience and
machine learning. Recent researches emphasized the biological plausibility of
Linear-Nonlinear-Poisson (LNP) neuron model. We show that with neurally
plausible settings, the whole network is capable of representing any Boltzmann
machine and performing a semi-stochastic Bayesian inference algorithm lying
between Gibbs sampling and variational inference.
Description
Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On
Boltzmann Machines
%0 Generic
%1 shao2012linearnonlinearpoisson
%A Shao, Louis Yuanlong
%D 2012
%K Linear-Nonlinear-Poisson Networks Neural models probabilistic
%T Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On
Boltzmann Machines
%U http://arxiv.org/abs/1210.8442
%X One conjecture in both deep learning and classical connectionist viewpoint is
that the biological brain implements certain kinds of deep networks as its
back-end. However, to our knowledge, a detailed correspondence has not yet been
set up, which is important if we want to bridge between neuroscience and
machine learning. Recent researches emphasized the biological plausibility of
Linear-Nonlinear-Poisson (LNP) neuron model. We show that with neurally
plausible settings, the whole network is capable of representing any Boltzmann
machine and performing a semi-stochastic Bayesian inference algorithm lying
between Gibbs sampling and variational inference.
@misc{shao2012linearnonlinearpoisson,
abstract = {One conjecture in both deep learning and classical connectionist viewpoint is
that the biological brain implements certain kinds of deep networks as its
back-end. However, to our knowledge, a detailed correspondence has not yet been
set up, which is important if we want to bridge between neuroscience and
machine learning. Recent researches emphasized the biological plausibility of
Linear-Nonlinear-Poisson (LNP) neuron model. We show that with neurally
plausible settings, the whole network is capable of representing any Boltzmann
machine and performing a semi-stochastic Bayesian inference algorithm lying
between Gibbs sampling and variational inference.},
added-at = {2013-05-27T13:18:31.000+0200},
author = {Shao, Louis Yuanlong},
biburl = {https://www.bibsonomy.org/bibtex/246b4bfb1a7fe5eb90b3ae0435ceac683/sidyr},
description = {Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On
Boltzmann Machines},
interhash = {094ce765ba90893be28110b529d2ad4a},
intrahash = {46b4bfb1a7fe5eb90b3ae0435ceac683},
keywords = {Linear-Nonlinear-Poisson Networks Neural models probabilistic},
note = {cite arxiv:1210.8442Comment: Submitted to International Conference of Learning Representation (ICLR) 2013},
timestamp = {2013-05-27T13:18:31.000+0200},
title = {Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On
Boltzmann Machines},
url = {http://arxiv.org/abs/1210.8442},
year = 2012
}