Scribd
Upload
25 views

Backpropagation Algorithm: Example: Prof. Navneet Goyal

The document provides an example of how the backpropagation algorithm works for updating weights in a neural network. It presents the forward and backward passes for a single training example, showing how the error is calculated and weights are updated using partial derivatives and the chain rule.

Uploaded by

harsh solanki
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
25 views

Backpropagation Algorithm: Example: Prof. Navneet Goyal

The document provides an example of how the backpropagation algorithm works for updating weights in a neural network. It presents the forward and backward passes for a single training example, showing how the error is calculated and weights are updated using partial derivatives and the chain rule.

Uploaded by

harsh solanki
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 19

Backpropagation Algorithm:

Example
PROF. NAVNEET GOYAL

Source: A Step by Step Backpropagation Example by Matt Mazur


Example
The Network
Example
single training example: given
inputs 0.05 and 0.10, we want the
neural network to output 0.01 and
0.99.
Example
The Forward Pass
Example: Forward Pass
Example: The Error
Example: The Backward Pass
Consider w5.
We want to know how much a change in w5 affects the total error!

By applying the chain rule we know that:


Example: The Backward Pass
Example: The Backward Pass
Example: The Backward Pass
How much does the output of change with respect to its total net input?

(remember – derivative of sigmoidal function!!)


Example: The Backward Pass
Finally, how much does the total net input of o1 change with
respect to w5?

Combining the three components, we get:


Example: The Backward Pass

Note the sign of the derivative wrt w5 (it is positive)


Example: The Backward Pass
We can repeat this process for updating w6, w7, & w8
Hidden Layer - updation of w1, w2, w3, & w4
Hidden Layer - updation of w1, w2, w3, & w4
Hidden Layer - updation of w1, w2, w3, & w4

Following the same process, we get

Check this out!


Hidden Layer - updation of w1, w2, w3, & w4
Hidden Layer - updation of w1, w2, w3, & w4

Similarly,
Final Points
o We’ve updated all of our weights!
o When we fed forward the 0.05 and 0.1 inputs originally,
the error on the network was 0.298371109.
o After this first round of backpropagation, the total error is
now down to 0.291027924
o After repeating this process 10,000 times, for example, the
error plummets to 0.0000351085.
o 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).

You might also like