The Mark I Perceptron was a landmark achievement in machine learning and it remains an iconic symbol of neural networks. At the same time, the operations of the machine itself are poorly understood. This thesis describes the Mark I, drawing from a variety of published and previously obscure sources. The Mark I was highly dependent on human interaction for its operations. To fully explain the Mark I's operations also requires recovering a plausible description of the distributed cognitive system that surrounded the Mark I while it was used for experiments at the Cornell Aeronautical Laboratory. Modern machine learning systems are largely autonomous unless they have been specifically designed for interactivity. The Mark I required human operators to function, but it was it was also designed leverage this interactivity so that researchers could explore novel techniques for training neural networks. A study of the interfaces and procedures of the Mark I provides useful interpretive tools to understand modern artificial intelligence and machine learning systems.
Copyright is held by the author.
This thesis may be printed or downloaded for non-commercial research and scholarly purposes.
Supervisor or Senior Supervisor
Thesis advisor: DiPaola, Steve
Member of collection