Teaching Machine Learning to Computer Science Preservice Teachers

Machine learning is a fast-growing field with various applications in artificial intelligence and data science. Recently, a new machine learning program have been integrated into the Israeli high school computer science curriculum and thus we added a new machine learning module to the Methods of Teaching Computer Science (MTCS) course, which is part of the teachers' preparation program. This machine learning module provides us a unique opportunity to teach both pedagogy and content with the same subject matter. After teaching the basics of machine learning, we asked the students to find similarities between human learning theories and machine learning algorithms. Students identified several interesting parallels: (a) Supervised learning is similar to behavioral learning as the machine learns to connect training examples (stimuli) with labels (behavior). Also, the learning is based on minimizing error (punishment) function, (b) Reinforcement learning is similar to behavioral learning as learning is based on feedback from the environment, (c) Constructivism can be identified in the iterative convergence of the algorithm; the inner model improves each iteration based on the current knowledge, and (d) Social learning is reflected in clustering as each cluster affects the learning of the other clusters. In our talk, we present the idea that computational mental models may be used to reinforce pedagogical mental models and vice versa.