Machine Learning // Avatars and Abstract Human Forms

Earlier this year when I was visiting the Barbican for the AI: More Than Human exhibition, there one piece of work that was stuck with me for a long time (and it was free). It was the interactive installation Future You by Universal Everything. Situated right in the main corridor on the Silk street entrance, the artwork is two pieces of massive screen that contains abstract geometries that follows your own movements, tracked by cameras on top. The sensing device kept track of locations and movements of your torso and limbs, and translated into the movement of a growingly complex geometric object.

It was an interesting experience observing people of all ages and genders moving and dancing in front of the screen and engrossed in looking at the mirror movement. When I couldn’t help but to give myself a chance to dance in front of it, I was immediately thrown in the world between me and the screen. There seemed to be a synthesis going on where the lifeless form in front of me became a part of my body because it followed my brain’s command seamlessly and abstract and basic geometry somehow obtained life-like features.

Margo K. Apostolos published back in 1990 a research journal on Robot Choreography, in which she examined the perspectives of both the artists and the scientists on the topic of technology aided art form – especially one where body is treated as a medium. In her proposal, robot arms, where certain levels of agilities were achieved for utilitarian purposes, can use such agility as a medium that represent or imitate a human body to express emotions the same way that we use our bodies as expression naturally. While she pinned down a list of parameters, such as the soft lines of movements, timing of movements and speed, the use of music and synchronization, she essentially defined a new medium of art in a methodical way.

In recent practice, James Leach and Scott Delahunta collaborated with the forward-thinking choreographer Wayne McGregor and dancers in his company to explore deeper implications of extending body as a medium to the digital and more abstract versions, which they called an Artificial Choreographic Entity. One of the main challenges, as they described, was “the lack of a commonly comprehensible description of what choreographic problem-solving or thinking was.” As part of the research, they interview a number of dancers to discuss what “the body” is, and many responses suggested that the relational aspect of the body, such as how bodies are perceived and generates responses from other bodies, is what makes up the qualities of the bodies. “You cannot be in the same space as another body and not feel a response.” McGregor said.

While technologies allow us to generate human like forms in all aspect – material(AI sex dolls), facial features(, voice(speech synthesis), and we’re closer than ever to the stage where we can create extremely human like objects, I remain particularly fascinated by these non-human-like avatars or mechanical objects that moves in a way that human beings can relate. It’s perhaps because of the drastic contrast between the differences of forms and the similarities of the motions and the deeper emotions that can be evoked. This makes me feel like I can extend myself in an unlimited amount of forms of my imaginations.

This experience became a driving focus among my interests in AI and embodiment. What’s more interesting to me is another project by Universal Everything Called Hype Cycle: Machine Learning, where videos were shown the process of using machine learning algorithms to teach these abstract geometries to move in a human like way. (The video shown here on the left is one of them.) When I watched the geometric objects move alongside with a human dancer, with patterns of movements they learned from human motions, I felt a strongly moved by the dance piece. This led me to think that perhaps something formally simple can arouse our sense of empathy and connections, or other emotional responses that we generally experience towards human agents.

““Choreography, the art of making dances, uses dance as a series of thythmic motions in time and space to express ideas through movement. The process of choreographing for a robotic arm combines a logical approach with a sensuous approach in a blend of artistic and scientific creativity.”

—Margo K. Apostolos”Dance+Becoming+Knowledge

  1. Universal Everything. ‘Future You’. Accessed 20 November 2019.

  2. Universal Everything. ‘Hype Cycle: Machine Learning’. Accessed 20 November 2019.

  3. Apostolos, Margo K. ‘Robot Choreography: Moving in a New Direction’. Leonardo 23, no. 1 (1990): 25.

  4. Leach, James, and Scott deLahunta. ‘Dance Becoming Knowledge: Designing a Digital “Body”’. Leonardo 50, no. 5 (October 2017): 461–67.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Create a website or blog at