Torcat Tiquacrat Mytics Ahtrae, 2018, Blum and Poe, Los Angeles
Scott Benzel, Torcat Tiquacrat Mytics Ahtrae Mvmt I, 2018 excerpt
Scott Benzel, Torcat Tiqaucrat Mytics Ahtrae
Acid Free, Blum and Poe, Los Angeles, May 5, 2018, 4PM
The generative model can be thought of as analogous to a team of counterfeiters, trying to produce fake currency and use it without detection, while the discriminative model is analogous to the police, trying to detect the counterfeit currency. Competition in this game drives both teams to improve their methods until the counterfeits are indistiguishable from the genuine articles.
Generative Adverserial Nets, Ian J. Goodfellow, et. al., arXiv:1406.2661v1, code and hyperparameters: http://www.github.com/goodfeli/adversarial
Maxout networks learn not just the relationship between hidden units, but also the activation function of each hidden unit.
Maxout networks, Goodfellow, Ian J, et al.. arXiv:1302.4389
Torcat Tiqaucrat Mytics Ahtrae, is candygrammar/chrome selected for via GAN (Generative Adversarial Network) from the titles of the two books on offer, Sonatine Bureacratique and Torch of the Mystics, LACA, 2018, with a significant remainder word redacted.
The performance is refuctored from the texts of Sonatine Bureacratique and Torch of the Mystics, the score and marginalia of Erik Satie’s Sonatine Bureaucratique and Cahiers d’un mammifère, Man Ray’s Enigma of Isadore Ducasse and Poem, Picabia’s SAVON and Thermomètre Rimbaud, and a Q modified Cicada and structured thusly:
- text, from the above sources via DCGAN (Deep Convolutional Generative Adversarial Network) and Maxout MLP (modifiedStackGAN), read by the author, as well as readymade poetry AIs.
- images generated via DCGAN trained from the above sources, projected or flatscreened.
- score generated from the above sources via P22 ‘johncagesys’, WaveGAN, Noatkl, IanniX, et al. interpreted by live electroacoustic ensemble, flute, viola, bass, and live digital processing.
Performance time variable.