UNSW
⃝c Alan Blair, 2013-20
UNSW
⃝c Alan Blair, 2013-20
COMP9444
Neural Networks and Deep Learning
Course Web Page
COMP9444 20T2
Overview
2
COMP9444 20T2
Overview 3
Lecture / Lab Schedule
Lectures
1a. Overview
https://www.cse.unsw.edu.au/~cs9444/20T2/
https://webcms3.cse.unsw.edu.au/COMP9444/20T2/
Online Lectures (Weeks 1-5, 7-10)
◮ Monday 5pm-7pm and Tuesday 5pm-7pm
Students are required to watch pre-recorded lecture videos before each session. The scheduled class time will take the form of an interactive video chat session, and will be used to briefly summarise the content, deliver additional material, and to answer any questions that you may have about each topic.
Labs (Optional, tentative) ◮ Wed 2-4 (Weeks 1-10) ◮ Thu 4-6 (Weeks 1-10)
As well as watching the lecture videos, consider doing these things:
UNSW
⃝c Alan Blair, 2013-20
UNSW
⃝c Alan Blair, 2013-20
COMP9444 20T2 Overview 1
Lecturer-in-Charge
Alan Blair
blair@cse.unsw.edu.au
K17-412C
9385-7131
◮ review the lecture material before and after the scheduled class ◮ discuss the material with fellow students if possible
◮ read up on the topics covered in each lecture
◮ complete relevant assignments and exercises, if any
◮ explore the topic by writing and running your own programs ◮ ask questions in an online consultation session
COMP9444 20T2
Overview 4
COMP9444 20T2 Overview 5
Textbook
Assumed Knowledge
The textbook for this course is:
Deep Learning
By Ian Goodfellow, Yoshua Bengio and Aaron Courville MIT Press
http://www.deeplearningbook.org https://mitpress.mit.edu/books/deep-learning
The course will assume knowledge of the following mathematical topics:
UNSW
⃝c Alan Blair, 2013-20
UNSW
⃝c Alan Blair, 2013-20
COMP9444 20T2 Overview
6 COMP9444 20T2 Overview
7
Planned Schedule (Weeks 1-5)
Planned Schedule (Weeks 7-10)
Week 1, Mon: Week 1, Tue: Week 2, Mon: Week 2, Tue: Week 3, Mon: Week 3, Tue: Week 4, Mon: Week 4, Tue: Week 5, Mon: Week 5, Tue:
Neuroanatomy, Perceptrons Backpropagation
(Labor Day Holiday) Probability, Backprop Variations Hidden Unit Dynamics
(1.2, 9.10) (4.3, 5.1-5)
Week 6: Week 7, Mon: Week 7, Tue: Week 8, Mon: Week 8, Tue: Week 9, Mon: Week 9, Tue: Week 10, Mon:
(Flexibility Week)
Reinforcement Learning
Deep Reinforcement Learning Hopfield Network, Boltzmann Machine Autoencoders
Generative Adversarial Networks Extension Topics
Review
(12.5.1.1)
(18.1, 20.9)
(16.7, 17.4, 18.2, 20.1-4) (14.1-5, 20.10.3) (20.10.4)
UNSW
⃝c Alan Blair, 2013-20
UNSW
⃝c Alan Blair, 2013-20
PyTorch
Convolutional Networks
Image Processing
Recurrent Networks, LSTM and GRU Language Processing
(7.9, 9.1-5)
(7.4, 8.4, 8.7.1) (10.2, 10.7, 10.10) (10.4, 12.4)
(3.1-14, 6.1-5) (7.11-12, 8.2-3)
Linear Algebra (2.1-2.8)
Probability (3.1-3.14)
Calculus and Chain Rule (6.5.2)
Students should study the relevant sections of the textbook (shown in brackets) and, if necessary, try to revise these topics on their own during the first few weeks of the course.
COMP9444 20T2
Overview 8
COMP9444 20T2 Overview 9
Assessment
Pytorch
Assessment will consist of:
Assignment 1 30% Assignment 2 30% Final Exam 40%
Please try to install Pytorch on your own laptop, and try to match the environment on the CSE Lab machines as closely as possible:
The assignments may involve, for example:
torch numpy sklearn
1.2.0 1.16.2 0.20.2
using code written in pytorch
writing your own code
running experiments and analysing the results
Further details will be provided on the course Web site. UNSW
⃝c Alan Blair, 2013-20
UNSW
⃝c Alan Blair, 2013-20
COMP9444 20T2
Overview 10
COMP9444 20T2
Overview 11
Plagiarism
Related Courses
PlagiarismistakenseriouslybyUNSW/CSEandtreatedasAcademic Misconduct. ALL work submitted for assessment must be your own work.
COMP3411/9414 Artificial Intelligence
COMP9417 Machine Learning and Data Mining
COMP9418 Advanced Topics in Statistical Machine Learning
COMP4418 Knowledge Representation and Reasoning
COMP3431 Robotic Software Architecture
COMP9517 Machine Vision
4th Year Thesis topics
Foranindividualassignment,collaborativeworkintheformof“think tanking” is encouraged, but students are not allowed to derive code together as a group during such discussions. In the case of a group assignment, code must not be obtained from outside the group.
Plagiarism detection software may be used on submitted work.
Academic Integrity and Plagiarism:
UNSW
⃝c Alan Blair, 2013-20
UNSW
⃝c Alan Blair, 2013-20
https://student.unsw.edu.au/plagiarism
python3 3.7.3