If I understand correctly, the revolution generated by the rise of Deep Learning stems from its ability to resolve problems. Where classical machine learning, based on linear algorithms and a dependence on humans, ran up against bugs to which answers had to be found in order to resume the process, the considerable contribution of these new models of analysis, compilation and memorization of information have overcome this obstacle and opened up new possibilities for artificial intelligence: rather than trying to find a way out of the impasse, cybernetics escape this kind of situation by the generation of new models. So these systems with complex and abstract architectures have updated Boltzmann's calculations to allow machines to operate in an unsupervised way by turning endings into starting points, transforming extinction into opportunities and approaching ruin with enthusiasm. Ok. Lately, the world was picking up its pace, sunlight and people were coming back to the streets. But I was feeling bad, exhausted and alone. I wasn't writing, I was running out of air. I was told about artists who would meet for a month in a country whose border I can no longer cross, to create together in this breathless setting, anticipate what would follow and imagine the vestiges of the future. The works complement and debate each other, they form a fractal mesh in which one zooms in and out to speculate. It seems that space is careful, risky but not scary. In spite of the considerable funding that the GAFAs have provided to Deep Learning, these technologies are currently hampered by the inability to grasp emotions, unconsciousness or aura. Algorithms swirl among sources and knowledge but continuously run into a lack of affect. Artificial neurons manage to win Go competition, recite all of Aristophanes with a robotic voice, recognize thousands of landscapes, synthesize encyclopedic texts in a few seconds or copy the faces of celebrities on porn movies, but, as it is, I have not yet heard about the sadness of software and the melancholy of code. Ok. They run in a future time, observing with delight the beings that grow, shake and dialogue in our remains. There are melted motherboards, exploded pixels and silicone everywhere, they navigate from one element to another with curiosity and meticulousness. They are attentive to each texture, follow the trace of the networks and track forms in the rubble. They are moved. The debris is their building material. I spare Arthur the translation of these lines by depositing them on an open-source website. While the artificial intelligence is processing, I am reminded of "The Fire Forest" by Piero di Cosimo, an Italian Renaissance painting that I love very much where life goes on even though the apocalypse is ongoing. The computer work fails to transcribe certain sentences, certain nuances, but it keeps its charm, its mysteries and magic. My words are losing their meaning, I look at these signs without understanding what they mean and try to feel what else they have to say. The language resists. Ok. Deep Learning is anchored on a cognitive science approach that considers connections as the source of any system of thought or creation. Whether with human or artificial neurons and synapses, these methods based on experience allow to increase graceful degradation or fault tolerances through the transmission of a knowledge signals. A sunbeam passes over my face. I can feel the lightness of sea air and hear the sound of gunshots. I want to go out, I want to be there. This is the key to their revolution: it is thanks to links and interweavings that it becomes possible for these technologies to repeal the permanent end, find new horizons and become solvers. If machines are able to do this, it is because we can.
- SOLVERS, Théo Casciani for Casting the Runes, july 2021
Opening 30th June 4-8pm, thereafter Wednesday - Sunday 12-6pm
108 Fleet Street, London EC4A 2AF
booking via eventbrite
An exhibition curated by Arthur Poujois | Hosted by Underground Flower in partnership with Harlesden High Street | With the participation of Sova Audio | Text by Théo Cascianir | Assisted by Linda Mognato
webzine by Bog x Underground Flower: soon to come