Machine Learning on Google Cloud
- 6 courses
- 18 hours
This path teaches course participants how to write distributed machine learning models that scale in TensorFlow, scale out the training of those models, and offer high-performance predictions. You will learn to convert raw data to features in a way that allows ML to learn important characteristics from the data and bring human insight to bear on the problem. You will also learn how to incorporate the right mix of parameters that yields accurate, generalized models and knowledge of the theory to solve specific types of ML problems. You will experiment with end-to-end ML, starting from building an ML-focused strategy and progressing into model training, optimization, and productionalization with hands-on labs using Google Cloud Platform.
Courses in this path
This section builds upon the previous course in the path. Google thinks about machine learning as being about logic, rather than just data. You will learn why such a framing is useful for data scientists when thinking about building a pipeline of machine learning models. You will learn about the five phases of converting a candidate use case to be driven by machine learning, and why it is important the phases not be skipped. You will examine how machine learning can amplify bias. Next you will examine why neural networks today perform so well in a variety of data science problems. You will then learn to set up a supervised learning problem and find a good solution using gradient descent. To do so, you must ensure that your datasets permit generalization; the course covers methods of doing so in a repeatable way that supports experimentation.
In this section you work with low-level TensorFlow, progressing through the necessary concepts and APIs so as to be able to write distributed machine learning models. You will learn, given a TensorFlow model, how to scale out the training of that model and offer high-performance predictions using Cloud Machine Learning Engine. You will learn how to perform feature engineering, including recognizing the elements of good versus bad features, and how you can preprocess and transform them for optimal use in your machine learning models. You will get hands-on practice choosing features and preprocessing them inside of Google Cloud Platform with interactive labs.
In this section, you will first learn about aspects of machine learning that require some intuition, good judgment, and experimentation. We call it the art of ML. We will learn the many knobs and levers involved in training a model. You will manually adjust them to see their effects on model performance. Once familiar with the knobs and levers, otherwise known as hyper parameters, you will learn how to tune them in an automatic way using Cloud AI Platform on Google Cloud Platform. You'll further spice things up by pinch of science, and the science involved in training neural networks. You will also learn the important concept of embeddings which is all about representing discrete objects such as words as real valued vectors. Finally, you will switch from working with TensorFlow's pre-built estimators such as DNN Regressor to building your own custom estimators