Are you a player or an observer?
Kill the process
When people tap on the box, it will reply to the sound sequence by tapping back from within
Spinoff of OutComes, this project came out as a way to analyze people’s reactions to the installation: they just wanted to play.
Kill the process an artificial neural network capable of learning and communicating with the audience, taking advantage of the articulation and subjectivity of rhythmic patterns as a language.
With each new received sequence, the system not only learn more rhythmic patterns (which thicken its database / vocabulary), but also studies the audience reaction to its response; understanding the relationships between the given sequences, looking for the syntax of the language the audience is defining.
Although at first glance it could be seen as another gadget dealing with the future of technology in our lives, it deals with a far more human question:
- Would you accept being excluded in the interest of improving the results? or when over time you become expendable,
- Would you force the process to start over from the very beginning? Even though you would be excluded again and again…
It emerges better prepared, faster, and vastly more precise than us… You can just listen and be an expectator of its evolution, or you can wipe out its memory so you are again the center of attention in a system that needs you to learn.
In its early stages the box utilizes the audience through the rhythms they contribute (by tapping it) to learn a language. A communication process arises, becoming more complex and interesting the more the audience interact with it.
Slowly, it starts to navigate the database finding new relations based on what people did most, and slightly modifying the records at the same time to avoid recursivity, evolving into something new with every iteration.
As the system develops it will try to chain a few sequences, offering a more elaborate response, responding to itself, to the point it generates something that doesn’t know how to continue so it needs a human input again. But as time passes by and the database gets richer, that self-response process will enter in continuous generation.
It will play a sequence according to what a human would most likely do, based on the information previously collected during the learning process, analyzes its own sequence and replies to it again and again. But at the same time each sequence is altered when played, purging the database on each iteration, avoiding to enter in an infinite loop.
It becomes an autonomous entity that finds its rhythmic capabilities more accurate than ours, so it excludes us from the equation, starting a self recognition process in which the outcome is completely uncertain, but will be modelled on what people did previously.
It just does what it does best: to learn & predict.
At the end of the day it is as dumb or as clever as the prediction algorithms from Google that helps it end your sentence before you type it, that helps you find what you are looking for even if you don’t know how it is called; it is a machine, behaving as a human would do.