Syllabus

 

Lecture Slides:

Lecture 1: Machine Learning Basics.      pdf     video-1     video-2

Lecture 2: Feed Forward Neural Networks.      pdf    video-1     video-2

Lecture 3: Regularization Strategies For Deep Models.      pdf    video-1    video-2

Lecture 4: Optimization For Training Deep Models.      pdf-1    video-1   pdf-2   video-2

Lecture 5: Recurrent Neural Networks.      pdf-1    video-1   pdf-2    video-2

Lecture 6: Convolutional Neural Networks.      pdf    video

Lecture 7: Practical Considerations For Training Deep Models.      pdf    video

 

Project:

Project Description

Project Report Template

 

Announcement 1:

• Project Proposals are due on May 28th.
• Each team will be responsible to submit one proposal.
• Submission will be via email to aharakeh@uwaterloo.ca with a CC to stevenw@uwaterloo.ca.

 

Announcement 2:

• Project Milestones are due on July 4th.
• Each team will be responsible to submit one report.
• For the report format check bellow.
• Submission will be via email to aharakeh@uwaterloo.ca with a CC to stevenw@uwaterloo.ca.

 

Announcement 3:

• Project Reports are due on August 10th.
• Each team will be responsible to submit one report.
• For the report format check bellow.
• Submission will be via email to aharakeh@uwaterloo.ca with a CC to stevenw@uwaterloo.ca.

 

Presentations:

 

10/May/2017:

Jungwook Lee :

3D Object Proposals using Stereo Imagery for Accurate Object Class Detection.

presentation

Mathew Angus :

MultiNet: Real-time Joint Semantic Reasoning for Autonomous Driving.

presentation

 

17/May/2017:

Melissa Mozifian :

Deep Sliding Shapes for Amodal 3D Object Detection in RGB-D images.

presentation

Samin Khan :

Efficient Deep Models for Monocular Road Segmentation.

presentation

 

31/May/2017:

Jason Ku :

Multi-View 3D Object Detection Network for Autonomous Driving.

presentation

 

7/June/2017:

Sean Walsh :

We don’t need no bounding-boxes: Training object class detectors using only human verification.

presentation

 

14/June/2017:

Martin Cote :

Fully convolutional networks for semantic segmentation.

presentation

Video

Adam Sanderson:

Deep MANTA: A Coarse-to-ne Many-Task Network for joint 2D and 3D vehicle analysis from monocular image.

presentation

Video (Second half, starts at ~17 minutes)

 

21/June/2017:

Mohammad Elbalkani :

SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation.

 

Resources:

Deep Learning Book

Stanford’s CS 231n : Convolutional Neural Networks For Visual Recognition

1. Course
2. Google Cloud Getting Started Tutorial
3. Google Cloud GPU Tutorial
4. Amazon Web Service Tutorial

Stanford’s CS 229 : Introduction to Machine Learning

For any comments or inquiries about the course, please contact Ali Harakeh: aharakeh[at]uwaterloo[dot]ca.