In 1951, Marvin Minsky built one of the first artificial neural networks, known as the SNARC (Stochastic Neural Analog Reinforcement Calculator), as part of his doctoral research at Princeton University. This early network was designed to simulate a rat navigating a maze, using randomly connected tubes filled with liquid mercury to represent synapses. Although rudimentary by today’s standards, the SNARC was a foundational step in neural network research, demonstrating the potential of using physical systems to mimic cognitive processes.