Welcome to the art and science of machine learning. In this data science course you will learn the essential skills of ML intuition, good judgment and experimentation to finely tune and optimize your ML models for the best performance.
In this course you will learn the many knobs and levers involved in training a model. You will first manually adjust them to see their effects on model performance. Once familiar with the knobs and levers, otherwise known as hyperparameters, you will learn how to tune them in an automatic way using Cloud Machine Learning Engine on Google Cloud Platform.
Course overview highlighting the key objectives and modules. First, you will learn about aspects of Machine Learning that require some intuition, good judgment and experimentation. We call it the Art of ML. You will learn the many knobs and levers involved in training a model. You will manually adjust them to see their effects on model performance.
The Art of ML
In this course you will learn about The Art of Machine Learning. We will review aspects of machine learning that require intuition, judgment and experimentation to find the right balance and what’s good enough (spoiler alert: it's never perfect!).
In this module you will learn how to differentiate between parameters and hyperparameters. Then we’ll discuss traditional grid search approach and learn how to think beyond it with smarter algorithms. Finally you’ll learn how Cloud ML engine makes it so convenient to automate hyperparameter tuning.
A pinch of science
In this module, we will start to introduce the science along with the art of machine learning. We’re first going to talk about how to perform regularization for sparsity so that we can have simpler, more concise models. Then we’re going to talk about logistic regression and learning how to determine performance.
The science of neural networks
In this module we will now be diving deep into the science, specifically with neural networks.
In this module, you will learn how to use embeddings to manage sparse data, to make machine learning models that use sparse data consume less memory and train faster. Embeddings are also a way to do dimensionality reduction, and in that way, make models simpler and more generalizable.
In this module we will go beyond using canned estimators by writing a custom estimator. By writing a custom estimator, you will be able to gain greater control over the model function itself.
Review the key concepts we covered in the Art and Science of ML course.