The big hype these days is how machine learning is going to take away everyone’s job in a few years. Robots will be everywhere to do everything. Computers will drive our cars, do our chores, and we are stuck in a life of leisure. Alpha Go recently beat the best player in the world and there seems to be no stopping those smart computers. To understand where us humans stand we should dive into the machine algorithms used for game playing. The history of game playing with machine learning is reviewed to see where we were and perhaps where we are going. We dive into machine learning algorithms used to play checkers, chess, tic-tac-toe, backgammon, chess, Jeopardy! and Go. Then we look closer at the neural networks, reinforcement learning, and Monte-Carlo Tree Search—algorithms used to teach Alpha Go to play at a high level. To illustrate the Alpha Go approach a Java application I wrote which uses self-learning playing to learn to play a game of tic-tac-toe without losing will be demonstrated. We can then stand back and look at what is machine intelligence and compare weak and strong AI. Finally, we address those robots who are coming to take your job away. What jobs are at risk? What is the recipe to create a robot programmer using machine learning? What skills will you need to succeed in a world of robot programmers?
Bio Richard Abbuhl:
I am a developer, technical lead, consultant, and architect with more than 25 years of experience working both in the U.S. and in Europe. I have worked on numerous projects for small to large-sized companies and have developed software tools and enterprise applications for finance, medical, logistics, business services, and database mining. I also have an MBA from a U.S. AACSB-certified school which I hope can help companies build better products and provide better services with the goal of increasing quality, customer satisfaction, and ultimately higher profitability.