HAPSAT Graduate Workshops - First Fall Session

HAPSAT GRADUATE WORKSHOPS

The First Neural Network

In-person event | Wednesday, Oct. 1, 2025, from 12:00pm-1:30pm

VC303

Abstract: Modern AI is based on an adaptive algorithmic architecture: neural networks. Although nowadays they are studied almost exclusively by computer scientists and have gained an independence of their own, they once represented the convergence of several lines of inquiry from neuroanatomy and psychology to physics and computing. Aided by mid-twentieth-century computing, neural networks were the testing grounds of psychological tenets, philosophical doctrines, and neurophysiological mechanisms, simulating their commitments on a substrate thought to be shared with our very own nervous system: neurons and synapses. Yet, the plausibility of how neural networks abstracted these biological entities has always been questioned, growing into a marked divide between the fancifully imagined artificial and the empirically grounded biological. Here, I turn to a time before the rise of such prickly disciplinary boundaries, when neural networks were first developed, and I ask the following questions: what are neural networks? what makes them “neural”? These questions might seem quotidian, easily answered through a quick search—and perhaps they are if we are interested in knowing about modern neural networks. Yet, if the aim is to explain the unifying concept behind neural networks throughout history, then modern definitions and their genealogies do not suffice. Instead, we are forced to question the very foundations of the term and how it is distinguished from similar ideas. I answer these questions through the proxy of identifying the first neural network. The title of first is not a matter of handing out trophies, but serves as a heuristic to navigate candidate conceptions of neural networks and precisely disambiguate them from other theoretical constructs. Ultimately, my journey backwards in time suggests that neural networks precede, both historically and conceptually, the neurons and synapses that purportedly constitute them. If so, what are neural networks networks of? The nineteenth-century answer I reach has a lot to do with what makes neurons and synapses neural in the first place.

Graduate Workshop Organizer: Rebecca Muscant