Methods Outline

Requirements

You should now have a good idea of how you will complete your project. Create a new page (likely using your previous submissions as a starting point) and add a “methods” section.

• Revise your introduction and related works.
• Fill in a more complete version of your methods section with a five to ten sentence outline. Incorporate relevant information from your project updates. Relevant information includes the
  • software you are using,
  • datasets you are using,
  • tools you will use for analysis,
  • possible pitfalls,
  • etc.

This doesn’t need to be perfect, but this will provide a good chance for feedback.

Once this new section is fully filled-in (at the end of the semester), it will describe how you developed your project.

Here is an example (and I’ve added comments in parentheses that would not be part of the submission):

“We will create our own dataset by scrapping images from bing.com. (And then later you can fill in details about how many images and what techniques were used for scrapping).”

“We hope to collect 100 thousand images, but if we are unable to we will search for a similar already-existing dataset. (Collecting data is hard, so it is good to have a backup plan.)”

“For our model, we are using a pretrained convolutional neural network from TensorFlow Hub. (Later you would specify the chosen model and give a reasoning for the choice.)”

“Our project will focus on finding the best classification accuracy possible. (This might change later, but it gives your team something on which you can focus and code.)”

“We will use Gradio to build a simple web application around our trained model. (Maybe you get this done quickly and decide to build a full web application using the ITS web servers.)”

You will submit a link to your web-page on gradescope.

All of the information you provide is tentative, and I expect many groups to change their minds as projects evolve.