• L6: Machine learning - unsupervised learning and reinforcement learning

    In this module, you will learn how computers can be taught to learn "actively," enabling them to independently discover interesting relationships in data.

    For some tasks or problems, it is not known what the correct label or optimal decision is, for example, when companies want to automatically divide their customers into groups in order to develop targeted marketing strategies. In this module, you will learn how machines can use unsupervised learning to recognize patterns in data without a “correct answer” (i.e., labels) being specified. Unsupervised learning algorithms enable computers to recognize structures and relationships in unlabeled data. In this module, you will also learn about the reinforcement learning approach. This approach is inspired by behaviorist learning theories and allows computers to gradually make optimal decisions through rewards and punishments.

    In summary, this module covers the two other fundamental approaches to machine learning: unsupervised learning and reinforcement learning.

    In this module, you will learn ...

    • to differentiate between the problems of supervised, unsupervised and reinforcement machine learning.
    • how unsupervised learning algorithms recognize clusters and identify outliers.
    • how agents make increasingly better decisions during reinforcement learning through trial and error.

    • discuss the social implications of recommendation systems.

     
     

    We strongly recommend that you work through Module 06 on a computer in order to make optimum use of the learning resources.