The Born Perceptron: Mimicking Quantum Mechanics and Distributed Measurements in Time
Aman Chawla
Abstract
This paper develops a neural architecture capable of reproducing several structural aspects of quantum mechanics. The proposed model, termed the Born Perceptron, extends the classical perceptron by structuring outputs as value–index pairs and interpreting normalized activations as probability distributions over indexed outcomes. By allowing indices to inhabit geometric manifolds rather than discrete labels, the perceptron naturally represents probabilistic sampling over state spaces. When stacked sequentially, such perceptrons generate stochastic trajectories reminiscent of quantum state evolution under repeated measurement. Introducing complex-valued activations yields probabilities determined by squared magnitudes, reproducing the Born rule. The framework is further extended to treat measurements distributed in time, particularly position samplings. In this regime, the neural system produces sequences of sampled positions whose joint probabilities resemble those derived in path-integral formulations of quantum mechanics. The resulting architecture offers a classical computational model that captures several conceptual features of quantum theory including measurement, collapse, trajectory evolution, and amplitude-based probability rules.