Karaoke of Dreams

These days in the world of tech we see Machine Learning as the new catchall solution to social and technical problems. From the project that is working on translating the speech patterns of whales to facial recognition algorithms, to image processing that will finally let us "enhance" images, just like in the movies. But, Machine learning contains a fundamental flaw, an achilles heel, which is that it is the projection of our own human minds, a manifestation of a wish--of how we want to see reality. This has been demonstrated by projects such as Memo Akten's "Learning to See" and studies on confirmation bias, and prediction technologies. Karaoke of Dreams creates a space where a participant, a human singing karaoke, can experience what it feels like to perform in a totalizing environment created by Machine Learning. The song lyrics, music, harmonies, and stage lighting are all generated and leave nothing for the human to do but sing the song. In this action, we the viewers get to witness the very empathetic, emotional, and personal experience that is Karaoke, and the singer gets to perform it, while also experiencing it themselves. Karaoke of Dreams is an ML Karaoke that generates songs and video based on user input of song titles. In this case we asked for members of brahman.ai to send in titles on our Discord server. Karaoke is a profoundly soul-bearing activity where the singer pours their heart out in song. Karaoke of Dreams is a multi-dimensional machine learning matrix operating in the aural dimension on harmony, melody, lyrics and style. In the visual dimension it is operating on images tied to words using attention GANs. In the linguistic dimension it is using GPT2 fine tuned on pop song lyrics. In the spatial dimension it is a 4 fold rotational tesseract, outlining 3D space. The human element is your voice and performance of the generated concert experience. The Karaoke of Dreams envelops the singer like an AI womb, a safe, warm, pulsating center allowing the singer to pour their heart out.The decision not to reveal the mechanisms of the ML used in generating these titles or have any kind of user-feedback loop, or development cycle was made in an effort to leave the emotional experience of the karaoke private to the singer and onlookers to experience. We try to use tech today to quantify all the minutest, and most intimate aspects of human exprience. This project puts the human into a space of reckoning with smart technology, and leaves the outcome, or the judgement something for the human to ponder on for themselves.