I AM ALIVE by Stanza. Artificial intelligence machine learning artworks

I AM ALIVE 2023

An AI artwork which is alive and growing, learning from data sets over time to feed itself and evolve.

The entire orientation of the machine world is mediated by numbers. For the machine, reality is binary— on and off. Any knowledge about the world that we learn from the machine goes through this process of abstraction. As we become more dependent on our machines, we need to understand the underlying limits and boundaries of this abstraction.

The artwork is learning from a real-time agency of thousands of data feeds from the whole country using newly learned modified behaviours that have become alive. A neural network creates patterns inside patterns of data to reform the invisible across the whole of the country. The generative artwork is made using transport, weather, and pollution data in Germany using a newly created real-time STANZA.API with unsupervised machine learning processes built in. You see the re-imagined data from the whole country as a living system.

Data as Material. Processing the data in artistic form. This is a snapshot of the current state of the neural net on the server. The learning is cumulative, and each spawn of new areas represents a change in the neural net. The nodes are plugged into the real-time representation of the weather feed. If a change in the top level of the network it spawns a new node. An augmentation exists not only in the top layer of the neural net but the spaces in between where the data changes so what we are doing is interrogating the arrays which store the values on each of the nodes in the networks as well as interrogating the weights and strengths of each of the connections; so as the Stanza.API is updated it allows the integration of any layer in the net. What is finally created are visualisations of the maths processes of the internal nodes as the values resonate through the neural net. This interrogation of the real-time neural nets creates these visualisations and artworks.

Technology

The Stanza.API and database are written in node.js. The neural net is ported from C to JS. Using Adaptive Resonance Theory 2- [ART 2] to create unsupervised learning methods. As a result, it is adaptive and creates new clusters of behaviour; they are always able to learn new input patterns without forgetting the past. Unsupervised learning algorithms learn patterns from untagged data. The machine is forced to build a concise representation of its world and then generate imaginative content. The paths create emergent neural networks of data from newly implemented weather pollution transport feeds and present it all in JSON. The artwork is made with JavaScript. To display the work it's best on up to three projectors or three very large screen and please note you need an internet connection.

Data

Neural Nets. Can mix all the data feeds or run ART2 over any of the inputs.

Weather: Consists of country wide date including temperature, wind speed and direction, rain, snow, humidity and pressure.

Pollution. Station name, GPS, Time, Carbon Monoxide, Nitrous Oxide, Humidity, Ozone, Sulphur dioxide, Temp, Wind.

Transport Time. GPS, Trains Taxis, Buses, Trams.

Exhibitions.

Deutsches Museum Nuremberg Germany. 2023