COMP219: Advanced Artificial Intelligence (2019/2020)
Staff
Office Time
Every week, I will try to arrange a one-hour office time, but won't be a fixed time. So the following timeslot will be subject to change. However, please send an email to me to ensure my availability before coming.
Aims
- To equip students with the knowledge about basic algorithms that have been used to enable the AI agents to conduct the perception, inference, and planning tasks;
- To equip students with the knowledge about machine learning algorithms;
- To provide experience in applying basic AI algorithms to solve problems;
- To provide experience in applying machine learning algorithms to practical problems;
Learning Outcomes
- Ability to explain in detail how the techniques in the perceive-inference-action loop work
- Ability to choose, compare, and apply suitable basic learning algorithms to simple applications
- Ability to explain how deep neural networks are constructed and trained, and apply deep neural networks to work with large scale datasets
- Understand reinforcement learning, and is able to develop deep reinforcement learning algorithms for suitable applications
Lectures
There will be 26-30 lectures. The lecture notes will appear here one week ahead of the lectures.
-
Week 1
-
Week 2
-
Week 3
-
Week 4
-
Week 5
-
Week 6
-
Week 7
-
Week 8
-
Week 9
-
no lecture
-
no lecture
-
no lecture
-
Week 10
Tutorials:
Please follow the "Python Machine Learning" book to get hand-on experience. Please check VITAL for the learning materials. Every week, I will nominate a chapter or a section of the book as a tutorial for you to follow on.
-
Week 2: Chapter 1
-
Week 3: Chapter 2, the first two sections, up to page 32
-
Week 4: Chapter 3, Choosing a classification algorithm, first step with scikit-learn, and modeling class probabilities via logistic regression
-
Week 5: Chapter 3, maximum margin classification with support vector machine, and solving nonlinear problems using a kernel SVM
-
Week 6: Chapter 3, decision tree learning, and k-nearest neighbors
-
Week 7: Chapter 12
For the installation of tensorflow or other software package, it is recommended that you do this through virtual envionrments, which may be able to go around the user priviledge. See https://www.tensorflow.org/install/pip#2-create-a-virtual-environment-recommended
Also, if needed, please install Jupyter Notebook. Please read its document to understand what it is.
Assignments:
There will be two project assignments, which will be disclosed in due course. Each coursework assignment takes 10% of the final mark.
Note: this is different from last year!
TextBooks:
For general game design and implementation, the following books are recommended:
- Christopher Bishop, Pattern Recognition and Machine Learning, Springer-Verlag New York, 2006.
- Ian Goodfellow, Yoshua Bengio, and Aaron Courville. Deep Learning. MIT Press. 2017
- Sebastian Raschka. Python Machine Learning. Packt Publishing. 2016.