Advanced Deep Learning - Week 4


Course starts soon..


QUIZ


We will start now with a quiz based on the first week material

You have 6 minutes to answer the quiz.

The quiz link:
Quiz Link
It will be copied in Mattermost and in the Zoom chat.

Project Groups

We need to make the groups to start preparing the projects

Project Workflow

Try to agree on how to meet and work together

Tools and Programming Environment

  • Atom - Teletype: real time collaboration, local resources
  • Google Colab: delay in real time collaboration, google cloud resources
  • VSCode/ium: real time collaboration, local resources
  • Gitpod: still learning about real time collaboration, cloud environment, no installation
  • Many more..

Project Timelines

Starting today --> presenting in the end of June

In 2 weeks there is a checkpoint, where you discuss the state of the project

The state would be about dataset, objective, architecture and related work

Then you have 4 more weeks until the presentations, which will be focused on implementation

We can have an extra meeting to discuss if you have doubts

A visual example

03 May --> 10 May: preparation

10 May --> 17 May: preparation week 2 19 May: Checkpoint!

17 May --> 24 May: start programming

24 May --> 31 May: more programming

31 May --> 07 June: getting some results and fine-tuning

07 June --> 14 June: working towards good results, images and comparisons

14 June --> 22 June: preparing the final presentation 21 June: Presentation!

In case of need, we may skip the session of the 21st and present on the 28th of June

Quiz (10 Minutes)


  1. Is the transposed convolution the inverse of the convolution? What are its advantages and limitations?
  2. What are the main characteristics that differentiates YOLO form the previous approaches and makes it so fast? And what are its limitations?
  3. Why do we need non-max suppression and how can we be sure that we are not suppressing different objects? Can it handle different classes?

Homework

Bennet for the YOLO Assignment

Atul for the Semantic Segmentation Assignment

For the next week

  • Start preparing the project - template folder, repository, collaboration
  • Assignment on Face Recognition (Monfared?)
  • Assignment on Neural Style Transfer (Abhinav/Kwabena/Usman?)