hints_for_doe_project.docx

neural_networks.pptx

Unformatted Attachment Preview

HINTS FOR DOE PROJECT
a.
b.
c.
d.
Click on File
Click on New Python 3 notebook
Click on run time
Change Runtime type to GPU.
Copy the given codes and paste in new cells. Press shift Enter to run. Or click Run
button.
2. Run the codes as it is and treat this as your base experiment.
3. Our goal here is to use DOE to find the best combination of hyperparameters that gives
the highest validation accuracy in a Neural Network. Now, keep all other
hyperparameters as it is and change Kernel (filter) size Conv2D to 1, run the codes and
note the validation accuracy. Again, change the Kernel (filter) size Conv2D to 2, run the
codes and note the validation accuracy. Repeat this for all possible value of Kernel (filter)
size Conv2D and note validation accuracy for each value of Kernel (filter) size Conv2D.
Now plot the graph between different values of Kernel (filter) size Conv2D (x-axis) and
validation accuracies (y-axis). Looking at this graph choose two values of the Kernel
(filter) size Conv2D that denotes the highest difference in validation accuracy. (i.e. PICK
TWO LEVELS of this hyperparameter)
4. Repeat the same process for all the hyperparameters and find two levels for each
hyperparameter from graph.
5. Design your 2π experiment. (fractional factorial designs)
6. Collect data from your Neural network for single replicate analysis.
7. Find the significance of each hyperparameter and build a regression model. Test your
regression model. (The output of your regression model should be close to 50-60%)
8. Validate your model with neural network. Select the different combination of
hyperparameter values and find the validation accuracy using your regression model. Run
the same combination of hyperparameters in neural network and see the difference
between your result and the result from neural network. Explain the reason for the
difference in the result.
Examples
Everyday examples of NN
Everyday examples of NN (Contd.)
β’ FaceApp
Everyday examples of NN (Contd.)
β’ Autonomous driving
Everyday examples of NN (Contd.)
β’ Autonomous driving
Everyday examples of NN (Contd.)
β’ Medical image segmentation
Who do they do it?
A NN is nothing but function:
π¦1 , π¦2 , β¦ , π¦π = π π₯1 , π₯2 , β¦ , π₯π
π§ = π€1 π₯1 + π€2 π₯2 + β― + π€π π₯π + π
π¦=
1
1 + π βπ§
Basic Unit: Neuron
Activation functions
Deep neural network
Deep neural network
π§ = π€1 π₯1 + π€2 π₯2 + β― + π€π π₯π + π
π¦=
1
1 + π βπ§
Training a DNN
Convolutional neural network (CNN)
CNN (Contd.)
Hyperparameters optimization
The performance of a neural network depends on many hyperparameters
For example:
1) Activation function
2) Number of neurons
3) Number of layers
4) Filter size
5) Number of filters
Learning rate
http://www.benfrederickson.com/numericaloptimization/

attachment

## Calculate the price of your order

550 words
We'll send you the first draft for approval by September 11, 2018 at 10:52 AM
Total price:
\$26
The price is based on these factors:
Number of pages
Urgency
Basic features
• Free title page and bibliography
• Unlimited revisions
• Plagiarism-free guarantee
• Money-back guarantee
On-demand options
• Writerβs samples
• Part-by-part delivery
• Overnight delivery
• Copies of used sources
Paper format
• 275 words per page
• 12 pt Arial/Times New Roman
• Double line spacing
• Any citation style (APA, MLA, Chicago/Turabian, Harvard)

# Our guarantees

Delivering a high-quality product at a reasonable price is not enough anymore.
Thatβs why we have developed 5 beneficial guarantees that will make your experience with our service enjoyable, easy, and safe.

### Money-back guarantee

You have to be 100% sure of the quality of your product to give a money-back guarantee. This describes us perfectly. Make sure that this guarantee is totally transparent.

### Zero-plagiarism guarantee

Each paper is composed from scratch, according to your instructions. It is then checked by our plagiarism-detection software. There is no gap where plagiarism could squeeze in.

### Free-revision policy

Thanks to our free revisions, there is no way for you to be unsatisfied. We will work on your paper until you are completely happy with the result.