David Gandy and his XK120 charm London’s creative quarter
| How charity In Place Of War channels creativity in conflict zones
| Interior designer Joyce Wang shares the latest trends in luxury
| Panasonic Jaguar Racing’s most successful year in Formula E
| Meet Jaguar’s new design director Julian Thomson
The revolution has been
The revolution has been to understand how to write code that can learn. The notion is not that new – code that ‘learned’ how to play noughts and crosses was rst cooked up in the 1960s by Donald Michie – but the true power of machine learning has only recently hit the mainstream. What has changed is that there is now a rich digital environment in which the code can roam and learn. For instance, the huge number of digital images online has led to visual recognition software being able to distinguish a picture of a dog from a cat, something that top-down coding had failed to achieve. If code is learning, changing, developing, mutating, then at some point it might start to do things that surprise the person who wrote the original code. Suddenly, there is a possibility for the code to be creative – after all, one de nition of creativity is that it is the making of something that is new, surprising and has value. The top-down style of coding limited the surprise factor; now, if the code is mutating, it has the chance to surprise us. But novelty does not guarantee value. Value is very culturally, historically and personally speci c. I might write a poem of huge value to me personally, but it might be regarded as having little value in the wider world. This is where machine learning can really be a game changer – if we give it data to learn of things that we do value, then it might be able to identify the key markers in the data that allow it to contribute something that we similarly recognise has value. A team at Microsoft and Delft University of Technology got an algorithm to analyse 346 paintings by Rembrandt The rarest form of creativity is when something appears seemingly out of nowhere and learn what it is that makes a Rembrandt so special. It wasn’t just able to use this learning process to recognise a Rembrandt, but even produced a new painting that could pass itself o quite convincingly as in the Rembrandt school if not by Rembrandt himself. However, we don’t simply want pastiche, more of the same. We want innovation. Many people believe that this is impossible. If a machine has to operate within the con nes of a system that we understand, how can it break out of it and show us something new? One of the interesting by-products of trying to get code to be creative is that it pushes us as humans to try to understand what causes us to make a transformational move into the new. The cognitive scientist Margaret Boden identi ed three di erent sorts of creativity. The rst is exploratory creativity, in which someone takes the rules of the game and pushes them to the extreme. This is something that a computer is likely to excel at. Then there’s combinational creativity. This is where someone tries to create something new by synthesising two previously unrelated worlds. An example is fusion 60 / Jaguar Magazine
Tech cooking, the art of combining styles of cooking from two di erent cultures, or similar creative fusions in music, painting, architecture and even writing. What is exciting is that by understanding how this fusion can lead to innovation gives one a template for coding such creativity. The AI researcher Francois Pachet tried to capture this process in what he calls the Flow Machine. It works by analysing the underlying style of one genre, learns the rules, and applies them to a completely di erent data set. So, a machine could be ‘taught’ Schoenberg’s style of serialist music but then asked to play the blues in this style. As with all artistic experiments, the result » Jaguar Magazine / 61
JAGUAR MAGAZINE celebrates creativity in all its forms, with exclusive features that inspire sensory excitement, from seductive design to cutting-edge technology.
The latest issue features a range of inspiring people: from Luke Jennings, creator of Villanelle, one of the most interesting television characters in recent times, to Marcus Du Sautoy, who ponders whether artificial intelligence is on the brink of becoming creative. Out on the road, we visit the US to explore the foodie heaven of Portland in a Jaguar I-PACE, take a Jaguar XE to the south of France to get a photographer’s viewpoint of the charming town of Arles, and much more.