The Origin of Edge of Stability

Elon Litman · 2026

Read The Origin of Edge of Stability by Elon Litman (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-20446.txt.

Description

Litman explains why gradient descent on neural networks pushes the Hessian eigenvalue to the edge value 2 over the learning rate. He defines an edge coupling, a function over consecutive steps whose coefficient follows from the update rule. He derives a recurrence and a loss-change formula whose sums force curvature to the edge. Zeroing both coupling gradients sorts the fixed points from the period-two orbits. Near a fixed point, a half-amplitude function decides which directions support orbits and on which side of the rate they appear.

Catalog record

CREATORElon Litman
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE82228
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.20446

Browse the full library · About the Tower of Babel