Hi I’m Piero, a research scientist at Uber AI Labs, I work on machine learning and natural language. I’m originally from Bari, Italy, the capital of sunny and beautiful Apulia. I studied computer science at my hometown’s university and then traveled a lot around the globe during my PhD.
I’ve had a passion for computers and videogames since when I was a kid, when my father gave me a Commodore 64 as a present for my 4th birthday. I studied computer science in order to be able to program my own games but during my university years, I came across artificial intelligence and machine learning, and I was so fascinated that I changed my path. Part of the merit (or guilt?) goes to my professor Nicola Fanizzi who suggested reading “Godel, Escher, Bach”, an amazing “metaphorical fugue on minds and machines in the spirit of Lewis Carroll” by Douglas Hofstadter. It opened my mind and fostered my interest.
I moved to the US a couple years ago, living in New York at first and then moving to San Francisco when the startup I was working for was acquired by Uber. Before then I worked at IBM Watson and Yahoo. I’m here at Uber as it’s the place where I can pursue my interests, make a living out of it, and work with an amazing team of researchers.
I have many passions such as philosophy, Japanese and European graphic novels, robotics and art history but videogames and computation art have been my most visible ones. I was a videogame journalist while I was in college and used to collect old consoles and games.
I also developed a couple of little games, one of which “No! Birdie! No!” made it in the PC Gamer ‘Best 100 Web Games of all time’, even if I’m not really sure it deserved that much honor.
Now I mostly play indie games that I really appreciate rather than mainstream big budget titles.
Regarding my interest in computational art, it started from reading an essay by Umberto Eco, the combinatorics of creativity, where he basically says that nothing is original in itself, everything is a combination of something else and because of that the real important act is discerning what is interesting from what is not interesting.
Following this inspiration I teamed up with some skilled artists to try to build tools that enable this principle. The first outcome was a pixel art portrait editor called BitBuddies where you could select facial traits from a set of hand drawn ones (drawn by my friend Giuseppe Longo, a well known game artist in the Italian and British scene). You could combine them like you would do with the Mii in Nintendo consoles but you could also randomize them).
After teaching at XYZ, a designer summer school in 2014, I tried to follow the idea of generative / parametric design to geometric shapes coming from my research work. The result was Rosemantic, a piece of software written in Processing, where the user can specify ranges of variability of some parameters of the shapes and the software randomizes them returning a set of different shapes that are all coherent to each other because they are generated through the same process.
At that designer summer school, I also met Nicolò Loprieno, a street artist who was making murals based on geometric principles inspired by the logo and the architecture of the Casa da Musica in Oporto.
We decided to build software that could generate infinite variations of shapes according to his artistic styles and we came up with SkolpTiles. He currently uses it for coming up with new ideas, projecting them on walls and making new murals. The latest thing we did together is using these shapes as inputs for a wood 3d printing machine to make sculptures out of them that he then paints.
He recently sent me pictures of laser cut wood obtained from the files generated by my software. He painted it over with matte spray paint and made a wooden colorful sculpture out of it. I was really amazed, I love this combination of digital and analog and will probably try to explore it more in the future.
I believe having a creative outlet is really beneficial to me: it both helps me relax and helps me go in diffuse mode with my brain, when you don’t try to solve a problem, but rather let ideas flow, and that in the end also makes it possible to find solutions to problems you were not able to solve while focusing during work hours.
My latest project is an exploration of the use of Deep Learning for generating abstract art. It is inspired by Ken Stanley’s Compositional Pattern-Producing Network, but adds a little recursive twist to it. It can generate high resolution abstract images that can be played with changing some parameters.
I’m also currently working on a couple of ideas inspired by Piet Mondrian. One is mixing programming languages, generative art and machine learning (through evolutionary computing), while the other one is a more straightforward infinite parametric replicator of Mondrian’s art.
I love almost all art, but the kind of art that I find more intriguing is the art that explores geometry and patterns, possibly in combination with colour. The story behind it is a bit convoluted… I’m a big fan of Japanese manga and when I was a teen I was reading Kentaro Miura’s Berserk avidly. His graphic style was reminiscent of renaissance engravers like Dührer, but added a twist of impossible architectures and geometries like Escher did. So I discovered Escher’s work and fell in love with it and his use of visual paradox.
My tastes are aimed towards geometric simplicity and elegance, and they resonate a lot with the artists of the vanguards of the first half of the 20th century that I came across when I was 18. In particular architecture inspired art, constructivism and futurism became my favourite art styles, which included De Chirico, De Stijl with Piet Mondrian and the Bauhaus with Kandisky and Moholy-Nagy. Malevich is another of my favourites.
I see a connecting thread between those artists and their use of geometry and colour, the stunning simplicity and the absolute ideas they carry with them. those aspects make their art pieces somehow detached from space and time, even though they were influenced heavily by development in society and technology of their time, and will keep being relevant and influential in the future.
If you would like to learn more about Piero’s projects, check out www.w4nderlu.st