BusinessAugust 27, 2026
backpropagation
bak-prop-uh-GAY-shun
Definition
A foundational algorithm used to train neural networks by calculating the gradient of the error function and adjusting the internal weights of the model to reduce prediction errors.
Etymology
A compound of 'back' and 'propagation', the term originated in the 1970s and 1980s in computer science, referring to the way error signals are propagated backward through the layers of a neural network to update its parameters.
In the news
The article describes backpropagation as the process through which modern financial models refine their signals across diverse asset classes, enabling them to adapt to changing market conditions.
Deep Learning & Alpha Generation in Quant Finance 2026
Read the full article ↗Rebellion Research