Welcome back. I'm Andrew Zaldivar, a Data Scientist in the Research and Machine Intelligence group at Google. I work on developing the benefits of artificial intelligence and machine learning for everyone. And this is part of what I'll be talking about in this course. So far, we've talked about the strategy behind machine learning, about what machine learning means, what problems it can solve, and how to put it into practice at your company. Besides these technical and business aspects, another thing to consider is how fair and inclusive your models are treating your users. A key aspect in your machine learning strategy is to build machine learning systems in an inclusive way. So, in this module, I will show you how to identify the origins of bias in machine learning. And sometimes, it comes down to the training data itself. Then I will show you ways in which you can apply an inclusive lens throughout the machine learning development process, from the data exploration all the way to evaluating the performance of your training model. So, let's delve in. We'll first watch a video that explains where bias in machine learning originates, and the importance of building inclusive machine learning systems. After the video, I'll walk through some of the ways in which you can understand the trade-offs between the outcomes of your machine learning system and your users, and how these trade-offs map to evaluate the metrics that you can compute. From there, I'll introduce equality of opportunity, a methodology that builds on top of these evaluation metrics in order to achieve a more desirable outcome. An outcome where there is an equal chance of a machine learning system correctly classifying an outcome, irrespective of any sensitive attributes. And finally, as we know, machine learning systems are fueled by data. So, getting the best results out of a machine learning system requires that you truly understand your data, and that holds true for making machine learning systems inclusive. So, in this last section, I'll showcase an open-source visualization tool for machine learning data called Facets, which helps you explore the intricacies of your dataset, and provides some suggestions on what to look for when assessing the inclusiveness of your training data.