Special series If you enjoy our AI & ML conference MCubed, you might also like our brand new monthly webcast on all things machine learning from a software development perspective. There’s just one more week until episode one hits and drags you into the depths of data analytics.
Every developer working in the context of machine learning will sooner or later stumble upon the often cited “garbage in, garbage out” formula, which serves as a reminder that no good model can come from crappy input. But how can you tell good from bad data?
Knowing some expected patterns might be a good start and investigating them can help catch fraudulently generated data before you build your whole system on it. In episode one of our sparkling new MCubed web lecture series, Prof Mark Whitehorn will introduce you to one very common though often overlooked such pattern: Benford’s distribution.
Learn what it looks like, and why it occurs, so that you can decide whether it’s absence in the data you’re using is to be expected or rather a telltale sign that a closer look should be in order. We’re pretty sure you’ll leave the session seeing it everywhere – from the street you live in, to your local hardware store – and itching to examine some data to see if Benford’s distribution really is THAT common.
However, this hopefully won’t be the only thing this webcast will inspire you to do, as we’ll also take a look at current news in the practical machine learning space, such as tool updates, new machine learning services, or recently introduced projects. Like that, you’re up to date on latest developments and maybe find some things that can help you move your own work along.
Each event will feature a hand-selected expert talk in which practitioners will share insights from their daily work, answering questions such as how to know if a model is any good, what a solid production infrastructure looks like, and which tools are essential to implement ideas quickly. We’ll also use the chance to revisit some basics now and again, as having a solid foundational knowledge of a topic never hurt anyone.
You have a machine learning-related topic that you would like to know more about? Excellent! Let us know and we do our best to unpick it in one of the upcoming episodes.
Don’t forget to sign up here and we will see you on September 2. ®