by Girinandini SinghFeb 27, 2021
Digital artist Sofia Crespo creates vibrant, captivating work that teeters on the brink of pure recognition before promptly dissipating into total abstraction. Her pieces draw viewers in, presenting themselves as pictures of one or multiple forms of biotic life, yet, upon spending the slightest length of time viewing her work, this illusion is shattered and one realises that these are, in fact, several different images that have been mashed together seamlessly. The result is the creation of an image representative of an entity that is at once, reminiscent of easily recognisable plants or animals, yet, is altogether different; perhaps, alien even. Her work fits into the ambit of GAN (Generative Adversarial Networks) art, which is an image generative system that possesses a predisposition to creating ‘indeterminate’ images, as computer scholar Aaron Hertzman describes it in his crucial paper, “Visual Indeterminacy in GAN Art”. Hertzmann’s paper hypothesizes that this indeterminacy is a consequence of a powerful-but-imperfect image synthesis model that must combine general classes of objects, scenes, and textures, and, using Crespo’s work as an example, one may typify GAN art as being visually dense to the point of approaching psychedelia.
The artist was born in Argentina, but has lived in Germany for nearly a decade. She has studied Art Direction but does not feel that this is particularly relevant to the creation or reading of her work. Instead, it is likely a more appropriate and fruitful venture to study her growing oeuvre as a product of the intersecting interests and preoccupations of contemporary digital artists and machine learning professionals. She tells STIR, “I was lucky to get a scholarship at the School of Machines, Making & Make-Believe in a course about Machine Learning for Artists taught by Gene Kogan and Andreas Refsgaard, and that allowed me to learn various techniques which I afterwards continued to research on the internet. I would say the way I learn is largely through trial, error, and curiosity”.
Post-graduation, Crespo likely saw her fascination regarding machine learning intersect with an innate, perhaps undiminishable human interest in natural organisms, and this would eventually bring her to ponder the possible physiclaities and behaviours of artificial lifeforms as they may exist within the digital space. She mentions that it is interesting to consider the accuracy with which we can now extract the very concept of what life looks or acts like, and that this is founded not only within the sophistication of the image capturing, editing and reproduction software involved, but also within the critical relationship between human and machine. In fact, it is the ability of the human mind to learn and recognise visual patterns that allow for any meaningful engagement to be made in this regard, and Crespo’s work is aimed at disrupting and interrogating this very pattern recognition process in order to arrive at some crucial substrate that defines how we perceive life itself. She says, “Artificial life is something that inspires me, together with the idea that we can extract the concept of what 'life' looks like or acts like in a digital context. It blurs the lines between what we consider to be alive, and it certainly stimulates the areas of the brain that perform pattern recognition. Working on envisioning artificial life is an attempt to better understand why what we consider looks alive by trying to distil its 'naturalness' visually”.
Crespo cites computer researcher Karl Sims, scholar Bert Wang-Chak Chan and computer artist William Latham among her influences. She also references the Italian artist Luigi Serarfini’s seminal work, the Codex Seraphinianus, which is a starkly unique and fascinating tome that documents Serafini’s speculations on an imaginary world, populated by fantastical creatures, that, like Crespo’s hybrid forms, are themselves born of a biotic admixture. Crucially, while Serafini was limited by the tools of his time and had to contain his imaginings within the realm of illustration, Crespo has at her disposal a wealth of technology with which to shape her technique. However, this does not make her creative pursuit any less arduous, and she informs STIR, “There's not one single process that I use, but most of my experiments involve dataset creation. So, I spend lots of my time gathering datasets, either in 2D or in 3D. Sometimes, these are very large datasets of maybe 250000 images, and other times, it's as few as 20. The process also has a technical phase, where I try to figure out how to reach a result according to the concept behind a particular piece”. After this, Crespo must commit herself to a lengthy exploratory period that sees her go through hundreds and thousands of outputs in order to pick the ones that best fit the original concept, or perhaps even enhance or mutate it in some way.
Crespo’s artistry has been both aided and hindered by the ongoing pandemic. Prior to the outbreak, she was looking forward to travelling the world and exhibiting her work physically at various avenues. However, the extent of the virus’ damage along with the ensuing lockdowns created a precarious situation. She mentions that she was highly uncertain as to what trajectory her artistic career may be forced to take as a result. However, adaptability prevailed, and the present paradigm opened up new opportunities for her, pushing her deeper into the virtual realm along with other professionals who possessed similar interests. She says, “I am grateful for the collective effort to make art still be part of a conversation about what's happening globally. Some of the things we got to participate in were Nvidia's online GTC conference and Arebyte's Real-Time Constraints exhibition, which was a lot of fun because we worked in a team consisting entirely of artists and curators to design a plugin that would display art pop-ups”. Speaking of the future, Crespo seems quite secure in her ability to pursue her line of artistic enquiry, and mentions that she looks forward to collaborating with scientists and engineers in order to build new tools to engage with nature; ideally serving to educate and inspire people with regards to the beauty of biodiversity.