and here is a short guide about it…
Having barely passed term 1. It is essential to revisit all the chapters again and revise all the theory and concepts.
INTRO
Term 1 is called Foundation of AI, and the title pretty much summed it up. It is probably a revision for people with basic background in AI and/or programming. It starts with setting up your environments with Anaconda, Github and their project submission process, then you dive straight into your first project in programming a Sudoku solver. It was a relatively big jump and I felt that people with lesser experience programming in Python will have a hard time. I definitely struggled, especially having started 4 weeks late into the course, literally playing catch up all the way…
SOLVING SUDOKU
The Sudoku project was like a warm up to the rest of the AI course, it introduces you to functional programming and basic algorithm, as well as the importance in understanding of Object Oriented Programming (OOP) with Python. If you do not know or mastered the basic of Python OOP programming by the end of this chapter, you will have no chance of completing Term 1. I would therefore recommend anyone who is not familiar with Python to brush up on that before taking on this course.
EASY FOR SOME…
The next 7 chapters and 2 projects are literally all the basic computation theories related to machine learning and AI. If you have a BSc in Computer Science or a related engineering degree, these chapters will probably be a breeze for you. Here are the titles:
- Introduction to AI
- Introduction to game playing
- Advanced game playing
ADVERSARIAL SEARCH
After the above chapters, which involved many sub sections including theories and algorithms like Evaluation functions, MiniMax, AlphaBeta pruning, Probabilistic AlphaBeta and etc. You with then going on to your 3rd project...
Building an adversarial search agent. Algorithm like MiniMax and AlphaBeta are new to me, as I studied Engineering sometime ago or I simply did not pick the right modules. I had a hard time understanding how to write the algorithm to perform these functions. Luckily the tests provided was very helpful and informative in what the acceptance criterias are. It even provided an opportunity to have your agent competing with agents built by other. students.
RESEARCH REVIEW
There was also a requirement to write a short review with research paper related to AI. They have supplied 3 intriguing which were:
I went with AlphaGo as I have been following DeepMind’s development and their publications for some a while.
MORE THEORIES
By this time it may seems like the chapters before each project are preparations and requirements to complete the projects. However I was wrong…
- Search
- Simulated annealing
- Constraint satisfaction
- Logic and reasoning
Hint: Download the next project and beginning reading through it and give it a go, even if you have not started / completed any of the chapters. It will give you a head start and some understanding of what you are trying to achieve.
PLANNING
- Planning
Creating a Domain-Independent Planner. This project was more about understanding planning and it is a walk through of the planning chapter. It also involved reading one of the chapter from the book Artificial Intelligence: A Modern Approach by Norvig and Russell (Which they have provided the relevant chapter as a pdf, however I did buy the book which is a great reference to other topics.)
PROBABILITY AND BAYES THEOREM
The next 2 chapters are basic probabilities and introduction to Bayes’ Theorem. I felt more at ease with these topics, mainly because I read a whole book on Bayes Theorem (Bayes Theorem. An introduction for beginner). However it is just an introduction to the much in depth chapters to follow…
- Probability
- Bayes Net
MORE BAYES NET AND HMM
The following chapters are the real deal…
- Inference in Bayes Net
- Hidden Markov Model (HMM)
To be continued…
