Skip main navigation

Meet the expert: Rebecca Fiebrink

In this video, you’ll meet artist Rebecca Fiebrink, inventor of the Wekinator tool.
I’m really interested in how we can build better machine learning tools for musicians and artists and other creators. I’m interested in how we can teach machine learning to people who aren’t necessarily computer scientists. And I’m interested in understanding better what machine learning and AI have to offer to people in creative domains as well as other domains in which they want to do cool things. So I think there’s a growing recognition that machine learning and artificial intelligence are really powerful tools. But I think it would be a shame if the only people who really got to make use of those tools were computer scientists or people working at really big tech companies.
So the Wekinator is a piece of free and open source software that I’ve built. I started making it in about 2008. And at that point in time, I’d been using machine learning to do music analysis for a few years. I was working on some of the problems that underpin what we now know in services like Spotify where you might want to do music recommendation and say, OK, I’ve got this large music data set. I’m going to use machine learning to say something useful about what’s in it or what people might want to listen to. And I got really interested in what musicians might do with similar technology.
I’m a musician myself and I saw a lot of potential for people to use these kinds of techniques in live performance to build new types of instruments, to do more effective work with data in the context of composition. But I didn’t see very many people using them because at that point in time, it was really hard to use machine learning at all if you weren’t really well-versed in programming, if you didn’t have a really deep understanding of the algorithms and you didn’t build a lot of stuff from scratch yourself.
So machine learning is really useful for a lot of things, but it’s not a magic bullet. There are still some things that are really difficult to do. For instance, working with certain sensors, it might not always be obvious what that sensor is sensing. And even if you have a lot of signal processing expertise, maybe you have a PhD, you still might not exactly know how to take that data you’re getting from a sensor or from the world and how to best represent it when you give it to a machine learning algorithm. So that’s a challenge that I’ve been working on in my research for the last year or so.
There’s also still challenges around helping creators and other members of the general public understand what machine learning is really good for. If you want to build something with machine learning but you haven’t studied it before, you’re going to have questions about what it can and can’t do. And those questions can be really hard to answer.
So to do the work that I do, I draw on really a variety of very different skills. Obviously, I’ve had to develop a number of technical skills. I use programming a lot in my work. I like building software. It’s great to be able to build something myself and then just test it out really quickly rather than relying on someone else to build it. And of course, I have academic training in machine learning and related disciplines like mathematics. But that’s really not the end of the story. I would say there are two other sets of skills that are equally important in my work as an academic, and one of those is around research.
So the kind of research that I do often fits under the umbrella of something called human computer interaction. So I’m trying to figure out, hey, what is this technology good for, or what are the challenges that people encounter when they’re trying to use this technology in practise? And to do that, you need to have a set of skills around something called research methods. Part of that, for me, is around, how do you get information from people through things like interviews, through things like surveys? Part of it is just basic empathy and listening.
It’s being able to sit down with someone who might be the user of your tool or might be a potential user of something you’re making and trying to really figure out where they’re coming from. What is it that they do? What are their needs? What are the challenges that they encounter? And then thinking creatively about how you might match that to, say, the technical skills that you have.

As Rebecca explains in the video, the Wekinator tool has made a huge contribution to the proliferation and democratisation of AI in the creative space.

She tells us how making the tool open source has encouraged creativity and removed some of the barriers to entry for people looking to experiment and create with AI.

As you watch, think about her views on the impact her work has had and the problems the Wekinator tool was intended to address.

This article is from the free online

Introduction to Creative AI

Created by
FutureLearn - Learning For Life

Our purpose is to transform access to education.

We offer a diverse selection of courses from leading universities and cultural institutions from around the world. These are delivered one step at a time, and are accessible on mobile, tablet and desktop, so you can fit learning around your life.

We believe learning should be an enjoyable, social experience, so our courses offer the opportunity to discuss what you’re learning with others as you go, helping you make fresh discoveries and form new ideas.
You can unlock new opportunities with unlimited access to hundreds of online short courses for a year by subscribing to our Unlimited package. Build your knowledge with top universities and organisations.

Learn more about how FutureLearn is transforming access to education