Zobrazeno 1 - 1
of 1
pro vyhledávání: '"Cao, Daniel Yiming"'
We study gradient descent (GD) dynamics on logistic regression problems with large, constant step sizes. For linearly-separable data, it is known that GD converges to the minimizer with arbitrarily large step sizes, a property which no longer holds w
Externí odkaz:
http://arxiv.org/abs/2406.05033
