In order to explore the intricacies of the ChipAI technology, the artist proposes to look back at one of the earliest forms of embedded AI technology, the Mark I Perceptron, developed by Frank Rosenblatt in 1957 at Cornell University. Rosenblatt, a neurobiologist who was interested in how the eye of a fly acted as both the sensor and processing unit for the fly to flee in certain situations. The Mark I perceptron was both a sensor and processing unit built in hardware that could identify objects placed in front of it. Using the methodology of the ChipAI system as a means of inspiration for a novel construction of a light and sound based installation, the artist plans to try and bridge the gap between one of the earliest forms of AI with new cutting-edge advancements.

In Mark II Perceptron, Matthew Biederman challenges us to reflect on the current state of artificial intelligence and its technological past. In this residency, he worked with ChipAI Project, a project that aims to develop an energy-efficient technology using neuron-like nanoscale light sources and detectors capable of addressing the predicted future needs of Artificial Intelligence systems and computing processors. Biederman proposed to look at the trajectory of machine learning and artificial intelligence from one of the earliest iterations, the Mark I Perceptron developed by Frank Rosenblatt in 1957, to the current research project ChipAI, aiming to highlight their common features, namely to establish that both the Mark I Perceptron and ChipAI have their foundation in biologic processes. One of the Artist’s signatures is to use the tools and methods he is exposed to during the art-science residencies he has undertaken by integrating them directly into the artworks produced as a result. The Mark II Perceptron residency resulted in an interactive, light-based sculpture reflecting on the biomorphic genesis and continuing field of machine learning and artificial intelligence. As viewers approach the sculpture, they see a fractured reflection of themselves, which sets off a chain reaction through the artwork – a simulation of a neural network that is witnessed in real-time through the flickering halos of light from behind each mirror. Waves of flickering light will pass over the surface of the sculpture interfering with the reflections of the viewers. By using the viewer’s own reflection within the work, and their presence to affect the spikes and flickers, the ‘life’ or biologic processes that are sought to be mimicked in AI is conveyed. Each movement by the viewers sets off new spikes, but due to the way the model reacts, it is unique each time creating new and unexpected relationships.

Mid-term video

Final video

Read the final report Discover more on the blog