from《NTFs and CryptArt-Experiment》 Venue: From the solo exhibition “rhizomatiks_multiplex” (Museum of Contemporary Art Tokyo, 2021) https://www.mot-art-museum.jp/en/exhibitions/rhizomatiks/ This installation uses event data from Opensea from one day before and after 3/11 10am EST, when MetaKova won Beeple’s NFT at Christies for $69,346,250. It looks back on this highly symbolic event in CryptoArt using machine learning techniques (Convolutional Neural Network, t-SNE) and data visualization. from《NTFs and CryptArt-Experiment》 Venue: Solo Exhibition “rhizomatiks_multiplex” (Museum of Contemporary Art Tokyo, 2021) https://www.mot-art-museum.jp/en/exhibitions/rhizomatiks/ This installation uses event data from Opensea around 3/11 10am EST, when MetaKova won Beeple’s NFT at Christies for $69,346,250. We look back at this symbolic event in CryptoArt using machine learning techniques (Convolutional Neural Network, t-SNE) and data visualization.

Daito Manabe’s new work where he explores “brain” and “body” in the form of lecture performance. This prototype anticipates the future when dance can be generated by the brain decoding technology to express a dance piece without constraints of the real bodies such as gravity and the limits of the range of joints motion with noises and glitches generated by distracting thoughts in the brain.

Terminal Slam Music Video is a Diminished Reality / Mixed Reality music video in which all advertisements in the city are erased or replaced with Squarepusher-related ads. Daito Manabe created this MV, imagining that in the near future, as shown in this music video, AR/MR glasses will make it possible to freely rewrite cityscapes. Using machine learning technology, we performed: - Object Detection (identifying the positions of objects in an image) - Semantic Segmentation (labeling each pixel in an image and categorizing image regions according to their meaning) - Image Inpaint (image restoration) Most advertising billboards in the city change periodically, so with current technology it was difficult for machines to recognize them accurately; therefore, the billboard position data was created manually. In addition, mask information was generated from the analysis data and positional data, and post-processing was performed, such as applying glitch effects only to areas where people or advertisements exist, and applying optical camouflage effects. Furthermore, because the process of extracting advertisement regions could not be accomplished using existing machine learning technologies alone, manual operations were also carried out for elements that could not be realized with the technology available at present. We anticipate that in the near future, all of this will be done automatically.