https://wiki.alt-text.eu/index.php?action=history&feed=atom&title=Gradient_descentGradient descent - Revision history2026-09-15T06:36:54ZRevision history for this page on the wikiMediaWiki 1.45.3https://wiki.alt-text.eu/index.php?title=Gradient_descent&diff=44&oldid=previmported>ALT-TEXT: Import: AI terminology and people glossary2026-09-07T10:36:21Z<p>Import: AI terminology and people glossary</p>
<p><b>New page</b></p><div>== Gradient descent ==<br />
The core optimisation method used to train most AI models: an algorithm that repeatedly nudges a model's [[Parameters (AI model)|parameters]] in the direction that reduces its errors, measured by a [[Loss function|loss function]], until performance stops improving. It is called "descent" because it can be pictured as taking small steps downhill on an error landscape towards the lowest point. (See also: [[Backpropagation]], [[Loss function]])<br />
<br />
[[Category:Glossary]]<br />
[[Category:Artificial Intelligence]]<br />
[[Category:Machine Learning]]<br />
</div>imported>ALT-TEXT