Jingfeng Wu
@uuujf.bsky.social
Postdoc at Simons at UC Berkeley; alumnus of Johns Hopkins & Peking University; deep learning theory.
https://uuujf.github.io
https://uuujf.github.io
slides: uuujf.github.io/postdoc/wu20...
uuujf.github.io
September 26, 2025 at 3:49 AM
slides: uuujf.github.io/postdoc/wu20...
2/2 For regularized logistic regression (strongly cvx and smooth) with separable data, we show GD, with simply a large stepsize, can match Nesterov’s acceleration, among other cool results.
arxiv.org/abs/2506.02336
arxiv.org/abs/2506.02336
Large Stepsizes Accelerate Gradient Descent for Regularized Logistic Regression
We study gradient descent (GD) with a constant stepsize for $\ell_2$-regularized logistic regression with linearly separable data. Classical theory suggests small stepsizes to ensure monotonic reducti...
arxiv.org
June 4, 2025 at 6:55 PM
2/2 For regularized logistic regression (strongly cvx and smooth) with separable data, we show GD, with simply a large stepsize, can match Nesterov’s acceleration, among other cool results.
arxiv.org/abs/2506.02336
arxiv.org/abs/2506.02336
1/2 For the task of finding linear separator of a separable dataset with margin gamma, 1/gamma^2 steps suffice for adaptive GD with large stepsizes (applied to logistic loss). This is minimax optimal for first-order methods, and is impossible for GD with small stepsizes.
arxiv.org/abs/2504.04105
arxiv.org/abs/2504.04105
Minimax Optimal Convergence of Gradient Descent in Logistic Regression via Large and Adaptive Stepsizes
We study $\textit{gradient descent}$ (GD) for logistic regression on linearly separable data with stepsizes that adapt to the current risk, scaled by a constant hyperparameter $η$. We show that after ...
arxiv.org
June 4, 2025 at 6:55 PM
1/2 For the task of finding linear separator of a separable dataset with margin gamma, 1/gamma^2 steps suffice for adaptive GD with large stepsizes (applied to logistic loss). This is minimax optimal for first-order methods, and is impossible for GD with small stepsizes.
arxiv.org/abs/2504.04105
arxiv.org/abs/2504.04105