Welcome to the fourth course in this specialization. Let's see what we've covered so far and take a glimpse at what's coming ahead. Now we move on to how you can improve the accuracy of your ML models. Here we are going to address two key aspects; Feature engineering, and then understanding that art and science of ML. Feature engineering itself is often the longest and most difficult phase of building your ML project. It's in this feature engineering process where you start with your raw data, and you combine that with your own domain knowledge to ultimately create what we call features, that will make your machine learning algorithms work. So how do you know whether or not a data feel will be useful inside of your model? What are those rules of making a feature good? And how can you create new features to further enrich your data set? We'll answer these and get real hands on practice creating new features with TensorFlow, inside this course on future engineering. So let's get started.