Hochreiter's TU Munich thesis identifies the vanishing gradient problem

Identified the vanishing gradient problem in a thesis at the Technical University of Munich

Why it matters

The first to show that the learning signal vanishes as depth and sequence length grow. The problem definition behind LSTM and residual connections

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

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

Added in Week 35, 2026 · See that week

Nearby on the same axis

Backpropagation Applied to Handwritten Zip Code RecognitionPressApproximation by superpositions of a sigmoidal functionOfficialMultilayer feedforward networks are universal approximatorsPressLearning long-term dependencies with gradient descent is difficultPress