Abstract
Machine learning can be regarded as the core knowledge of artificial intelligence, and its applications are becoming more and more pervasive. Today, machine learning is becoming indispensable knowledge for computer science and engineering students. Current machine learning course relies heavily on large amount of “data” and abstract math as a pedagogical paradigm. Such an approach can be too abstract for beginners. We propose a new pedagogical model using robotics as an example. Specifically, we focus on the topics of reinforcement learning and develop Q robot. Current robot platforms available for research are either very expensive or power hungry. We develop a minimal and low cost Q-Robot with two degrees of freedom . Q-Robot is very affordable and can be assembled in less than an hour. The robot can learn to craw from scratch within less than 10 min using Q learning algorithm in reinforcement learning. It will become a pedagogical tool to provide hands on experience for engineering students.