You are on page 1of 10

5/1/2017 A Step by Step Backpropagation Example Matt Mazur

A Step by Step
BackpropagationExample
Matt Mazur
Background
Home
About
Backpropagation is a common method for training a neural network. There is no
Archives
shortage of papers online that attempt to explain how backpropagation works,
Contact
but few that include an example with actual numbers. This post is my attempt to
Now
explain how it works with a concrete example that folks can compare their own
Projects
calculations to in order to ensure they understand backpropagation correctly.

Follow via Email


If this kind of thing interests you, you should sign up for my newsletter where I
Enter your email address to follow
post about AI-related projects that Im working on.
this blog and receive notifications
of new posts by email.
Backpropagation in Python
Join 2,325 other followers
You can play around with a Python script that I wrote that implements the
Enter your email address
backpropagation algorithm in this Github repo.

Follow
Backpropagation Visualization
About
For an interactive visualization showing a neural network as it learns, check out
Hey there! I'm currently a data
my Neural Network visualization.
scientist Help Scout where I
wrangle data to gain insights into
our product and business. I also Additional Resources
built Lean Domain Search,
Preceden and many other If you find this tutorial useful and want to continue learning about neural
software products over the years.
networks and their applications, I highly recommend checking out Adrian
Rosebrocks excellent tutorial on Getting Started with Deep Learning and
Python.
Search
Overview
Follow me on Twitter
For this tutorial, were going to use a neural network with two inputs, two hidden
Tweets by @mhmazur neurons, two output neurons. Additionally, the hidden and output neurons will
include a bias.
Matt Mazur Retweeted

Luma Theodoro Heres the basic structure:


@lumatheodoro

Here you have a panel of


clocks, which together form a
clock

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 1/10
5/1/2017 A Step by Step Backpropagation Example Matt Mazur

23 Apr

Matt Mazur Retweeted

Bill Gates
@BillGates
Indian Prime Minister
@narendramodi put a
spotlight on a subject that
most of us would rather not
even think about: b-
gat.es/2qbs2Kk

25 Apr
In order to have some numbers to work with, here are the initial weights, the
biases, and training inputs/outputs:
Matt Mazur Retweeted

Jason Cohen
@asmartbear

Avoiding the trap of low-


knowledge, high-
confidence theories
blog.asmartbear.com/low-
fact-trap.

25 Apr

Matt Mazur
@mhmazur
A Simple CROSS
JOIN Example
mattmazur.com/2017/04/23/a- The goal of backpropagation is to optimize the weights so that the neural
s
network can learn how to correctly map arbitrary inputs to outputs.

For the rest of this tutorial were going to work with a single training set: given
inputs 0.05 and 0.10, we want the neural network to output 0.01 and 0.99.

The Forward Pass

23 Apr To begin, lets see what the neural network currently predicts given the weights
and biases above and inputs of 0.05 and 0.10. To do this well feed those inputs
Matt Mazur Retweeted
forward though the network.
Adam Kucharski
@AdamJKucharski
We figure out the total net input to each hidden layer neuron, squash the total
Survivorship bias
xkcd.com/1827/ net input using an activation function (here we use the logistic function), then
repeat the process with the output layer neurons.

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 2/10
5/1/2017 A Step by Step Backpropagation Example Matt Mazur

Total net input is also referred to as just net input by some sources.

Heres how we calculate the total net input for :

21 Apr We then squash it using the logistic function to get the output of :

Matt Mazur Retweeted

Tim Urban
@waitbutwhy
Carrying out the same process for we get:
It's finally here: the full story
on Neuralink. I knew the
future would be nuts but this
is a whole other level.
waitbutwhy.com/2017/04/neur We repeat this process for the output layer neurons, using the output from the
al
hidden layer neurons as inputs.

Heres the output for :

20 Apr

Embed View on Twitter

And carrying out the same process for we get:

Calculating the Total Error

We can now calculate the error for each output neuron using the squared error
function and sum them to get the total error:

Some sources refer to the target as the ideal and the output as the actual.

The is included so that exponent is cancelled when we differentiate later


on. The result is eventually multiplied by a learning rate anyway so it
doesnt matter that we introduce a constant here [1].

For example, the target output for is 0.01 but the neural network output
0.75136507, therefore its error is:

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 3/10
5/1/2017 A Step by Step Backpropagation Example Matt Mazur

Repeating this process for (remembering that the target is 0.99) we get:

The total error for the neural network is the sum of these errors:

The Backwards Pass

Our goal with backpropagation is to update each of the weights in the network
so that they cause the actual output to be closer the target output, thereby
minimizing the error for each output neuron and the network as a whole.

Output Layer

Consider . We want to know how much a change in affects the total error,
aka .

is read as the partial derivative of with respect to . You


can also say the gradient with respect to .

By applying the chain rule we know that:

Visually, heres what were doing:

We need to figure out each piece in this equation.

First, how much does the total error change with respect to the output?

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 4/10
5/1/2017 A Step by Step Backpropagation Example Matt Mazur

is sometimes expressed as

When we take the partial derivative of the total error with respect to ,
the quantity becomes zero because does not
affect it which means were taking the derivative of a constant which is
zero.

Next, how much does the output of change with respect to its total net input?

The partial derivative of the logistic function is the output multiplied by 1 minus
the output:

Finally, how much does the total net input of change with respect to ?

Putting it all together:

Youll often see this calculation combined in the form of the delta rule:

Alternatively, we have and which can be written as ,


aka (the Greek letter delta) aka the node delta. We can use this to
rewrite the calculation above:

Therefore:

Some sources extract the negative sign from so it would be written as:

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 5/10
5/1/2017 A Step by Step Backpropagation Example Matt Mazur

To decrease the error, we then subtract this value from the current weight
(optionally multiplied by some learning rate, eta, which well set to 0.5):

Some sources use (alpha) to represent the learning rate, others use
(eta), and others even use (epsilon).

We can repeat this process to get the new weights , , and :

We perform the actual updates in the neural network after we have the new
weights leading into the hidden layer neurons (ie, we use the original weights,
not the updated weights, when we continue the backpropagation algorithm
below).

Hidden Layer

Next, well continue the backwards pass by calculating new values for , ,
, and .

Big picture, heres what we need to figure out:

Visually:

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 6/10
5/1/2017 A Step by Step Backpropagation Example Matt Mazur

Were going to use a similar process as we did for the output layer, but slightly
different to account for the fact that the output of each hidden layer neuron
contributes to the output (and therefore error) of multiple output neurons. We
know that affects both and therefore the needs to take
into consideration its effect on the both output neurons:

Starting with :

We can calculate using values we calculated earlier:

And is equal to :

Plugging them in:

Following the same process for , we get:

Therefore:

Now that we have , we need to figure out and then for each
weight:

We calculate the partial derivative of the total net input to with respect to
the same as we did for the output neuron:

Putting it all together:

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 7/10
5/1/2017 A Step by Step Backpropagation Example Matt Mazur

You might also see this written as:

We can now update :

Repeating this for , , and

Finally, weve updated all of our weights! When we fed forward the 0.05 and 0.1
inputs originally, the error on the network was 0.298371109. After this first round
of backpropagation, the total error is now down to 0.291027924. It might not
seem like much, but after repeating this process 10,000 times, for example, the
error plummets to 0.000035085. At this point, when we feed forward 0.05 and
0.1, the two outputs neurons generate 0.015912196 (vs 0.01 target) and
0.984065734 (vs 0.99 target).

If youve made it this far and found any errors in any of the above or can think of
any ways to make it clearer for future readers, dont hesitate to drop me a note.
Thanks!

Share this:

Twitter Facebook 718

Like
49 bloggers like this.

Related

The State of Emergent Experimenting with a Emergent Mind #10


Mind Neural Network-based In "Emergent Mind"
In "Emergent Mind" Poker Bot
In "Poker Bot"

Posted on March 17, 2015 by Mazur. This entry was posted in Machine Learning and tagged ai,
backpropagation, machine learning, neural networks. Bookmark the permalink.

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 8/10
5/1/2017 A Step by Step Backpropagation Example Matt Mazur

Introducing ABTestCalculator.com, TetriNET Bot Source Code Published


an Open Source A/B Test on Github
Significance Calculator

392 thoughts on A Step by Step BackpropagationExample

Older Comments

yoba
April 16, 2017 at 5:05 am

Very comprehensive and useful example. Many thanks !

Reply

Himanshu
April 17, 2017 at 5:25 am

While calculating w1, shoul i use updated w5 or original w5?

Reply

MaviccPRP
April 18, 2017 at 2:39 am

Really good tutorial Matt. I finally had something like a aha moment ;-)

With the inspiration of this tutorial I translated this real number example into matrix
multiplication operations.

Reply

jerpint
April 18, 2017 at 1:13 pm

Hello, thanks for your explanation. Can you please clarify why the biases b1 and b2 are
the same for their entire
respective layers? Shouldnt we have a bias vector b of the same length as the number of
neurons per layer?

Reply

Paul Martin
April 18, 2017 at 2:39 pm

Like to print things out and read over. Seems like the print out starts omitting detail once
it hits the hidden layers. Odd. All in all this is quite helpful to understand the back
propagation.

Reply

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 9/10
5/1/2017 A Step by Step Backpropagation Example Matt Mazur

Ab
April 24, 2017 at 5:35 pm

Really appreciate this, thanks!

Reply

KG
April 25, 2017 at 7:41 pm

thank you very much best explaining i have ever seen

Reply

mahdi
April 25, 2017 at 7:52 pm

that was very helpful < thanx from a fighter in


mahadi army

Reply

Older Comments

Leave a Reply

Enter your comment here...

Blog at WordPress.com.

https://mattmazur.com/2015/03/17/a-step-by-step-backpropagation-example/ 10/10