![Shi Zheng, The Imitator Will Dance, 2020](https://artlogic-res.cloudinary.com/w_1600,h_1600,c_limit,f_auto,fl_lossy,q_auto/artlogicstorage/aikeart/images/view/aa6e822fe4655823c9f6ce9e7738b36aj/aike-shi-zheng-the-imitator-will-dance-2020.jpg)
![Shi Zheng, The Imitator Will Dance, 2020](https://artlogic-res.cloudinary.com/w_1600,h_1600,c_limit,f_auto,fl_lossy,q_auto/artlogicstorage/aikeart/images/view/1a8d49845577435622dd77a7b55efd18j/aike-shi-zheng-the-imitator-will-dance-2020.jpg)
Shi Zheng 中国, 1990
15‘14” loop
The Imitator Will Dance is an ongoing series project which explores the subject of neural networks and its
creative potential. The artist has worked with the Generative Adversarial Network (GAN), a type of
machine learning algorithm to create the piece. By imitating the input dataset, GAN creates images that
can not be distinguished from the real samples in “the latent space” after numerous training results and
iterations. Instead of the “natural / real” training results that other similar research normally obtains, the
artists attempts to create potential images which are more uncanny and obscure through further control of
data arrangement and parameter adjustment. The generated images appear to be a blurred dream,
moving slowly in the framework of time.
For the version on show, the choreography video of "C" is used as the dataset for training. The images
generated after a large number of iterations are not only the imitation of dancers’ postures, but also a re-
simulation of the dancing behavior consciousness. However, the machine will never restore the body
being imitated. As part of the installation, the wall drawing of a diagram has further illustrated and
contextualized the latent “body” in the data system.