Werbos's Harvard PhD thesis describes backpropagation for multilayer networks

Described error backpropagation training for multilayer neural networks in his Harvard PhD thesis

Why it matters

Regarded as the first description, predating the 1986 rediscovery. The reference point in the debate over priority for backpropagation

Source — Primary sources first; secondary reports are labelled “Press”

Pressen.wikipedia.orghttps://en.wikipedia.org/wiki/Paul_Werbos

Added in Week 35, 2026 · See that week

Nearby on the same axis

Minsky and Papert publish PerceptronsOfficialFukushima publishes the NeocognitronOfficialNeural networks and physical systems with emergent collective computational abilitiesGovernmentLearning representations by back-propagating errorsPaper