Introduction to Deep Learning for Visual Computing

Class Meeting



News


  • May 3 Papers
    Session is officially closed. All results will be announced via blackboard.
  • Apr 26 Papers
    Peer evaluation server will close on May 01, 2017 at 2:11PM. Make sure to complete all your evaluations by then.
  • Apr 26 Project
    Submission link for final project is up. Requirements are availble.
  • Apr 23 Project
    Schedule for project interviews are available.
  • Apr 21 Project
    Signups for interviews are available.
  • Apr 17 Project
    Final project reports are also due on May the 1st.
  • Apr 17 Finals
    The final exam will be conducted on May the 1st at BYAC 110 from 9:50AM to 11:40AM.
  • Apr 17 Project
    Final project interviews will be conducted on Monday (24 April) through Thursday (27 April).
  • Apr 6 Exam 2
    Midterm 2 grades are available at the blackboard. Exam statistics:

    Count : 59
    Minimum : 0
    Maximum : 12.00
    Mean : 5.01
    Median : 5.0
    Variance: 8.13
    Histogram

    If you want to have a look at your exams, I'll be carrying the answers to office hours for the next three office hours (until 17th April). Requests for checking the exams beyond these three office hours will not be encouraged.

    Remember, you will not be allowed to take a photograph, photocopy or leave with your answers when you are checking them. If you fail to return the exam back, your score will be switched to a zero.
  • Mar 30 Project
    Project mentors are announced. Visit the projects page for details.
  • Mar 30 Project
    Midterm project reports are due on April 7th. Submission link is now open.
  • Mar 19 Papers
    Schedule for paper discussions are announced.
  • Mar 4 Papers
    Singup groups are open for the paper discussions.
  • Mar 3 MP 6
    Miniproject 6 is announced. Due on 24th March 2017.
  • Mar 3 Note
    Chapter 5 of the book is now available. Passkey will be provided in class as usual.
  • Feb 17 MP 5
    Miniproject 5 is announced. Due on 17th March 2017.
  • Feb 23 Papers
    List of papers for paper discussions are announced.
  • Feb 18 MP 4
    Miniproject 4 deadline is extended. New deadline: 3rd March 2016.
  • Feb 17 MP 4
    Miniproject 4 is announced.
  • Feb 16 Note
    Chapter 4 of the book is now available. Passkey will be provided in class as usual.
  • Feb 16 MP 3
    Miniproject 2 grades are available at the blackboard.
  • Feb 16 MP 1
    The class's best solution of miniproject 1 (from Avinash Kaitha) is now added to the course repo. Congrats Avinash!
  • Feb 13 Exam 1
    Midterm 1 grades are available at the blackboard. Exam statistics:

    Count : 63
    Minimum : 3.75
    Maximum : 12.00
    Mean : 9.56
    Median : 10.00
    Variance: 4.65

    If you want to have a look at your exams, I'll be carrying the answers to office hours for the next three office hours (until 20th March). Requests for checking the exams beyond these three office hours will not be encouraged.

    Remember, you will not be allowed to take a photograph, photocopy or leave with your answers when you are checking them. If you fail to return the exam back, your score will be switched to a zero.
  • Feb 13 Project
    Project proposal submission link is now open.
  • Feb 13 MP 3
    Miniproject 3 submission link is now open.
  • Feb 11 Quiz 1
    Quiz 1 grades are available at the blackboard.
  • Feb 10 Lec 5
    Full audio of the class is available. The screencast of the toolbox tutorial is hosted at youtube, link to which is also provided.
  • Feb 10 MP3
    Miniproject 3 is announced.
  • Feb 9 Note
    There is a small bug in the derivation of one of the analytical solutions that one of the students in the class has found and has graciously posted a corrected version at the discussions board
  • Feb 6 MP 2
    Due to requests, the deadline for miniproject 2 is now extended. The new deadline is Tuesday, Feubuary 14th. A submission link is created on blackboard.
  • Feb 6 MP 1
    Miniproject 1 grades are available in blackboard. Since we are all statistics nerds, here are some statistics of the grades.
    Count : 63
    Minimum : 0
    Maximum : 4
    Mean : 3.79
    Median : 4
    Variance : 0.72
  • Feb 3 Lec 4
    Lecture materials including audios, slides and notes are posted.
  • Feb 2 Project
    Call for projects is now available in the projects tab of the course website. Project proposals are due in two weeks. Refer to the call for more details.
  • Feb 1 Jobs
    If you already took EEE 508 in Spring 2016 or earlier and are interested in being considered for a project on deep learning and video compression and are not currently working on a research project with another faculty member, please email your CV for consideration to karam@asu.edu
  • Feb 1 Note
    Chapter 3 of the book is now available. Passkey will be provided in class as usual.
  • Feb 1 MP 2
    Miniproject 2 is posted. Due on 10th Febuary 2017. Official announcement, demo and discussion on the project will be done in class on the 3rd.
  • Jan 27 Lec 3
    Lecture materials including audios, slides and notes are posted. Some slides are repeated twice, one of which was notes from doc cam and the other is my personal notes.
  • Jan 27 MP 1
    Blackboard link open for miniproject submission. Its is due one week from today.
  • Jan 20 Lec 2
    Lecture materials including audios, slides and notes are posted.
  • Jan 19 MP 1
    Miniproject 1 is posted. Due on 3rd Febuary 2017.
  • Jan 19 Note
    Chapter 1 and Chapter 2 of the book are now available.
  • Jan 19 Note
    Course now has its own github repo. Any course-related code will be posted here.
  • Jan 14 Note
    If you know matlab well-enough, numpy is not that difficult to learn. Refer the Numpy-Matlab cheatsheet for a comparitive study.
  • Jan 14 Note
    Numpy primer for those who are new to python and numpy. In the course, we will use python 2.7 and numpy a lot.
  • Jan 13 Lec 1
    Logistics and introduction to image representations.
  • Jan 11 Note
    Welcome to CSE 591.


Course Description


This course will introduce some basic paradigms for deep learning, with a focus on deep architectures using convolutional neural networks. Topics to be covered include:

  • Introduction to visual representation & fundamentals of machine learning
  • Neural networks & backpropagation
  • Optimization techniques for neural networks
  • General deep learning paradigms (CNN, auto-encoders, GANs etc.)
  • Modern convolutional neural networks
  • Software implementation of deep learning
  • Selected recent advances in deep learning

The course is intended to be interactive by requiring students to read the textbook or papers before each lecture, and the lecture time will be used to share insights, discuss pros and cons of any techniques to be studied in that lecture. There will be approximately 8 weeks of lectures by the instructors, and the rest will be on paper discussion and students will be required to make presentations. The students are required to understand all the papers presented, since there will be a final exam that will contain questions about the presented papers. Time permitting, the class may invite other faculty or senior PhD students who have relevant research work and recent publications on deep learning to give guest lectures.

The syllabus of the course can be found here. A tentative and updated plan for the course is also available here. Course now has its own github repo. Any course-related code will be posted here.

Text Books and Materials


There are no prescribed textbooks that is mandatorily required. The material for most of the course will come from the following book:

Other books that will also be used for references include:
  • Deep Learning by Ian Goodfellow, Aaron Courville and Yoshua Bengio
  • Neural Networks for Pattern Recognition, Christopher Bishop.
  • Online book: Neural Networks and Deep Learning by Michael Nielson

Being a graduate-level course on an up and coming topic, this course will also involve material from materials not published in books but in peer-reviewed conferences and journals.

Pre-requisites


Students should have good working knowledge of calculus, linear algebra and basic probability theory. It is highly recommended for a student to first take at least one machine learning class (e.g., CSE 569 or CSE 575 at ASU) before registering for this course. This will be a seminar course and will cover fairly advanced techniques and their applications without spending much time on reviewing the basics; a student who has not taken any graduate-level machine learning class will find it extremely difficult to succeed in this class. In addition, proficiency in programming in python is required for doing the course projects and/or mini project assignments.

License and Copyrights


© 2017 Ragav Venkatesan and Baoxin Li.

All the material posted in this webpage and on the course blackboard including lecture notes, lecture audios, homework, miniproject and project questions, and other materials are all intellectual property owned by the instrutors and other persons. You are not allowed to share, re-destribute, or host on public domain any of the material, either in part or full.