neuronal network

Artificial Light

Max Seeger (DE)

How does a neural network learn to see and how can we shape the perception of artificial intelligence? In a very experimental and fundamental manner, this project tries to study perception and recreation of light phenomena by artificial intelligence.

Nothing wild in particular

Jeremie Nuel (FR)

Consciousness, memory, reality and identity are regular themes in Philip K. Dick’s books. These concepts have fed into AI research since the seventies. This work uses Andrej Karpathy’s neural network, known as “recurrent,” to train a computer program, using as source a text or a corpus. Once trained, the code writes, character by character, a new text. The corpus that feeds the neuronal network shortens Philip K. Dick’s work (52 books). As a mirrored entity, strange and undisciplined, the computer program tries to write, at every iteration, step by step, an often-absurd story, and sometimes, a poetic one.