• Unit 3: Machine Learning and Its Importance

    A big part of AI is machine learning (ML) – the ability to see patterns in existing data sets and leverage those through advanced algorithms to make better decisions in complex environments. ML has myriad applications and is responsible for many of the apps all of us know and use every day.

    This unit describes the various types of ML models: supervised, unsupervised, and reinforcement learning. You will learn how these models are different and how they can be leveraged in creating advanced decision-making systems, with many examples of applications utilizing these methods to advantage.

    Completing this unit should take you approximately 3 hours.

    • Upon successful completion of this unit, you will be able to:

      • describe the importance of machine learning in today's business environments;
      • compare and contrast the different types of machine learning models, including supervised, unsupervised, and reinforcement learning; and
      • apply the definition of machine learning to identify the kind of learning models needed in different real-world problems.

    • 3.1: Learning in AI and Agents

      Machine Learning is an integral and vibrant part of AI. It has become a complete field of study on its own. Learning has been recognized as a key component of intelligence, no matter how you define it. Today, machine learning is all about using various statistical methods to discover patterns in past data to predict what might happen in the future or to understand certain deeper properties of the data. Rather than coding explicit instructions to mimic the expertise of humans (as traditional programming tries to do), machine learning looks at patterns in past data to glean patterns that lie therein to determine solutions to problems.

    • 3.2: Applications of ML in Neural Networks

      This section introduces you to neural networks (NNs), another rapidly growing ML sub-area. Using a simple building block called a "neuron" that does a simple computation, a neural network is designed into many different layers, with each layer getting its inputs from the previous layer and feeding its results to the next. The network parameters are gradually adjusted to map all the known inputs to the known outputs. Many types of NNs are now being applied in areas like image recognition, speech recognition, reading of MRI charts, and numerous other applications in different domains.