# Difference between revisions of "Feedforward backpropagation"

(→Model) |
|||

(5 intermediate revisions by 2 users not shown) | |||

Line 1: | Line 1: | ||

− | '''Feedforward backpropagation''' is an error-driven learning technique popularized in 1986 by David Rumelhart (1942-2011), an American psychologist, Geoffrey Hinton (1947-), a British informatician, and Ronald Williams, an American professor of computer science.<ref name=Rumelhart1986> | + | <translate> |

+ | |||

+ | '''Feedforward backpropagation''' is an error-driven learning technique popularized in 1986 by David Rumelhart (1942-2011), an American psychologist, Geoffrey Hinton (1947-), a British informatician, and Ronald Williams, an American professor of computer science.<ref name=Rumelhart1986>Rumelhart, David E.; Hinton, Geoffrey E.; Williams, Ronald J. (October 8, 1986). [http://www.cs.toronto.edu/~hinton/absps/naturebp.pdf "Learning representations by back-propagating errors"] <em>Nature</em> <strong>323</strong> (6088): 533–536</ref> It is a ''supervised'' learning technique, meaning that the desired outputs are known beforehand, and the task of the network is to learn to generate the desired outputs from the inputs. | ||

== Model == | == Model == | ||

Line 9: | Line 11: | ||

Given a set of k-dimensional inputs with values between 0 and 1 represented as a column vector: | Given a set of k-dimensional inputs with values between 0 and 1 represented as a column vector: | ||

− | + | [[File:Hebb1.png|center]] | |

and a nonlinear neuron with (initially random, uniformly distributed between -1 and 1) synaptic weights from the inputs: | and a nonlinear neuron with (initially random, uniformly distributed between -1 and 1) synaptic weights from the inputs: | ||

− | + | [[File:Hebb2.png|center]] | |

then the output of the neuron is defined as follows: | then the output of the neuron is defined as follows: | ||

− | + | [[File:FfBp1.png|center]] | |

− | where | + | where [[File:RBM3.png]] is a sigmoidal function. We will assume that the sigmoidal function is the simple logistic function: |

− | + | [[File:RBM4.png|center]] | |

This function has the useful property that | This function has the useful property that | ||

− | + | [[File:FfBp2.png|center]] | |

Feedforward backpropagation is typically applied to multiple layers of neurons, where the inputs are called the ''input layer'', the layer of neurons taking the inputs is called the ''hidden layer'', and the next layer of neurons taking their inputs from the outputs of the hidden layer is called the ''output layer''. There is no direct connectivity between the output layer and the input layer. | Feedforward backpropagation is typically applied to multiple layers of neurons, where the inputs are called the ''input layer'', the layer of neurons taking the inputs is called the ''hidden layer'', and the next layer of neurons taking their inputs from the outputs of the hidden layer is called the ''output layer''. There is no direct connectivity between the output layer and the input layer. | ||

− | If there are | + | If there are [[File:FfBp3.png]] inputs, [[File:FfBp4.png]] hidden neurons, and [[File:FfBp5.png]] output neurons, and the weights from inputs to hidden neurons are [[File:FfBp6.png]] (<em>i</em> being the input index and <em>j</em> being the hidden neuron index), and the weights from hidden neurons to output neurons are [[File:FfBp7.png]] (<em>i</em> being the hidden neuron index and <em>j</em> being the output neuron index), then the equations for the network are as follows: |

− | + | [[File:FfBp8.png|center]] | |

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | If the desired outputs for a given input vector are | + | If the desired outputs for a given input vector are [[File:FfBp9.png]], then the update rules for the weights are as follows: |

+ | [[File:FfBp10.png|center]] | ||

− | + | where η is some small learning rate, [[File:FfBp11.png]] is an error term for output neuron <em>j</em> and [[File:FfBp12.png]] is a ''backpropagated'' error term for hidden neuron <em>j</em>. | |

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | where | + | |

== Derivation == | == Derivation == | ||

Line 56: | Line 47: | ||

We first define an error term which is the cross-entropy of the output and target. We use cross-entropy because, in a sense, each output neuron represents a hypothesis about what the input represents, and the activation of the neuron represents a probability that the hypothesis is correct. | We first define an error term which is the cross-entropy of the output and target. We use cross-entropy because, in a sense, each output neuron represents a hypothesis about what the input represents, and the activation of the neuron represents a probability that the hypothesis is correct. | ||

− | + | [[File:FfBp13.png|center]] | |

The lower the cross entropy, the more accurately the network represents what needs to be learned. | The lower the cross entropy, the more accurately the network represents what needs to be learned. | ||

Line 62: | Line 53: | ||

Next, we determine how the error changes based on changes to an individual weight from hidden neuron to output neuron: | Next, we determine how the error changes based on changes to an individual weight from hidden neuron to output neuron: | ||

− | + | [[File:FfBp14.png|center]] | |

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | We then want to change | + | We then want to change [[File:FfBp7.png]] slightly in the direction which reduces <em>E</em>, that is, [[File:FfBp15.png]]. This is called ''gradient descent''. |

− | + | [[File:FfBp16.png|center]] | |

− | + | ||

− | + | ||

− | + | ||

− | + | ||

We do the same thing to find the update rule for the weights between input and hidden neurons: | We do the same thing to find the update rule for the weights between input and hidden neurons: | ||

− | + | [[File:FfBp17.png|center]] | |

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | + | ||

− | We then want to change | + | We then want to change [[File:FfBp6.png]] slightly in the direction which reduces <em>E</em>, that is, [[File:FfBp18.png]]: |

− | + | [[File:FfBp19.png|center]] | |

− | + | ||

− | + | ||

− | + | ||

− | + | ||

== Objections == | == Objections == | ||

− | While mathematically sound, the feedforward backpropagation algorithm has been called biologically implausible due to its requirements for neural connections to communicate backwards.<ref> | + | While mathematically sound, the feedforward backpropagation algorithm has been called biologically implausible due to its requirements for neural connections to communicate backwards.<ref><em>Backpropagation: Theory, Architectures, and Applications</em>. Chauvin, Yves; Rumelhart, David E. (1995). Lawrence Erlbaum Associates, Inc. ISBN 0805812598</ref> |

== References == | == References == | ||

<references/> | <references/> | ||

+ | |||

+ | [[Category: Neural computational models]] | ||

+ | |||

+ | </translate> |

## Latest revision as of 03:14, 24 June 2016

**Feedforward backpropagation** is an error-driven learning technique popularized in 1986 by David Rumelhart (1942-2011), an American psychologist, Geoffrey Hinton (1947-), a British informatician, and Ronald Williams, an American professor of computer science.^{[1]} It is a *supervised* learning technique, meaning that the desired outputs are known beforehand, and the task of the network is to learn to generate the desired outputs from the inputs.

## Contents

## Model

Given a set of k-dimensional inputs with values between 0 and 1 represented as a column vector:

and a nonlinear neuron with (initially random, uniformly distributed between -1 and 1) synaptic weights from the inputs:

then the output of the neuron is defined as follows:

This function has the useful property that

Feedforward backpropagation is typically applied to multiple layers of neurons, where the inputs are called the *input layer*, the layer of neurons taking the inputs is called the *hidden layer*, and the next layer of neurons taking their inputs from the outputs of the hidden layer is called the *output layer*. There is no direct connectivity between the output layer and the input layer.

*i*being the input index and

*j*being the hidden neuron index), and the weights from hidden neurons to output neurons are

*i*being the hidden neuron index and

*j*being the output neuron index), then the equations for the network are as follows:

*j*and

*backpropagated*error term for hidden neuron

*j*.

## Derivation

We first define an error term which is the cross-entropy of the output and target. We use cross-entropy because, in a sense, each output neuron represents a hypothesis about what the input represents, and the activation of the neuron represents a probability that the hypothesis is correct.

The lower the cross entropy, the more accurately the network represents what needs to be learned.

Next, we determine how the error changes based on changes to an individual weight from hidden neuron to output neuron:

*E*, that is,

*gradient descent*.

We do the same thing to find the update rule for the weights between input and hidden neurons:

*E*, that is,

## Objections

While mathematically sound, the feedforward backpropagation algorithm has been called biologically implausible due to its requirements for neural connections to communicate backwards.^{[2]}

## References

- ↑ Rumelhart, David E.; Hinton, Geoffrey E.; Williams, Ronald J. (October 8, 1986). "Learning representations by back-propagating errors"
*Nature***323**(6088): 533–536 - ↑
*Backpropagation: Theory, Architectures, and Applications*. Chauvin, Yves; Rumelhart, David E. (1995). Lawrence Erlbaum Associates, Inc. ISBN 0805812598