Preterrelational Experiments: 001. Physari-Algo Interlocution

Adriana Knouf / Boston, MA, USA — Nov 12, 2019

Image by Adriana Knouf
Care must extend beyond the human and envelop our machine learning algorithmic (»algo«) interlocutors. How should we care for them in their development, in spatial absence within data centers or other distant locations? How do we provide them with intriguing data sources or novel environments to learn in? How might we alter our encounters with other-than-humans to swerve us away from the need for semiotic meaning and towards new modes of engagement with our enchanted universe?
María Puig de la Bellacasa writes that »it is not difficult to see how cyborgs and other hybrid beings can be called to support the importance of care in more than human worlds« as it allows us to »extend meanings of caring out of expected normalized forms of kinship to embrace the unfamiliar«.
The project puts a Raspberry Pi with an algo (in this case a GAN) that has sensing and augmentation abilities, in relationship with a slime mold (Physarum polycephalum), all in an observable habitat. The algo on the Pi senses its environment and modulates its behavior (through light, temperature, voltage) in dialogue with the slime mold. Physarum’s dynamics adjusts the ways in which the algo learns, creating an encounter without teleology. Advanced machine learning research observes algos through online dashboards that chart their development towards more »perfect« recognition for humans. Here, instead, we encounter this preterrelation (a relation »beyond«) through an online presence that offers a different form of non-teleological care. Transmutating the data into a 24/7 stream of vibration, sound, and 3D surfaces, the online component senses human behavior through touch interactions, bringing such activity back to the Physari-algo system for further interlocution. This online presence exists for us to modulate ourselves towards the perspectives of other-than-human entities.

Find more contributions in the archive