The face of your voice 3D, from the verbal to the physiognomic
Contemporary life seems to be an endless game of data quantification, moving across different cultural domains. The former is an articulation of the digitalisation process, which feeds the current obsession with ‘machine learning’ strategies. The latter can spin these calculations in unexpected directions. ”The face of your voice 3D” by Frederik de Wilde in collaboration with Tae-Hyun Oh (MIT), uses data from a short audio recording of a person speaking to reconstruct their facial image. Trained through millions of Internet/YouTube videos, the neural network learns the correlation between a face and a voice, with age, gender and ethnicity. The reconstructed faces are then the conceptual mutations of the acquired data, showing the critical ‘right’ and ‘wrong’ directions that the machine autonomously takes.