How DID YOU GET INTO Machine LEARNING?
It's been a bit of a journey. I stumbled onto computer programming at about the age of eight, and was instantly enthralled by the idea of being able to have a computer simulate different worlds from a set of instructions. It wasn't long before I was teaching myself all sorts of different tricks and techniques, and started writing small games. I always wanted to push the boundaries of what my games could do, particularly in having them play out differently every time. From then I wanted to be a game developer, and so went on to study computer science and applied maths at Wits to follow this dream.
I then discovered that I found the challenge of making interesting and immersive characters in games to be more exciting than the rest of the game development process, and spent more and more time thinking about the algorithms behind this. I kept wanting to learn more and figured an MSc was the right thing to do. So I headed off to Edinburgh, and was completely blown away by the number of different courses around AI and machine learning. I also took a class on intelligent robotics (based on the coolness factor) with Dr Subramanian Ramamoorthy, and realised that this field combined ideas from every course I had enjoyed in both computer science and applied maths. I ended up doing both my MSc and PhD research with him, and have subsequently been fascinated by the idea of agents learning more about the environments in which they exist in order to keep making better decisions and improving their performance at different tasks.
I then discovered that I found the challenge of making interesting and immersive characters in games to be more exciting than the rest of the game development process, and spent more and more time thinking about the algorithms behind this. I kept wanting to learn more and figured an MSc was the right thing to do. So I headed off to Edinburgh, and was completely blown away by the number of different courses around AI and machine learning. I also took a class on intelligent robotics (based on the coolness factor) with Dr Subramanian Ramamoorthy, and realised that this field combined ideas from every course I had enjoyed in both computer science and applied maths. I ended up doing both my MSc and PhD research with him, and have subsequently been fascinated by the idea of agents learning more about the environments in which they exist in order to keep making better decisions and improving their performance at different tasks.
WhAT WILL YOU Be teaching?
I will be presenting the session on Reinforcement Learning, together with Vukosi Marivate. This is meant to be an introduction to the core principles of reinforcement learning. We will focus on fundamental topics such as MDPs, dynamic programming, policy/value iteration, and algorithms for learning Q (value) functions.
What advice would you give to those getting started in machine/deep learning?
It's a really exciting time to be entering the field! Take advantage of all the great resources that are available, and learn from everyone you can. It's often far more motivating, and fun, to work with other people, and become part of a community in that way. Don't be afraid to have grand ambitions, to explore and try new and crazy ideas!