paper-with-me

홈 › Papers

Stability Annealing Selects the Implicit Bias of Smoothed Sign Descent: A Rate-Indexed Barrier Path on Separable Data

2026-07-07 · Xiangwu Wang, Chengwei Cao, Yicheng Song, Ran Bi, Peilin Yu arxiv

Adaptive gradient methods can favor max-margin separators that differ from gradient descent, yet a fixed positive numerical stability constant eventually changes the update geometry again. This paper studies the rate-controlled middle case for full-batch linear classification on separable data. For memoryless stability-annealed smoothed-sign descent with weighted exponential loss, we prove that the normalized iterates converge to the minimizer of a convex Burg-type barrier over a margin slice. The proof rewrites the dynamics exactly as entropic mirror ascent on a concave dual objective, controls the dual gap by a KL recursion, and yields an explicit S_t^{-1/2} normalized-iterate envelope. The static barrier geometry is fully characterized, including KKT conditions and both endpoint limits. Experiments validate the exact dual identities to floating-point error, illustrate the predicted path and rate diagram, and show an empirical fixed-epsilon crossover scaling in cumulative time. We further report robustness and boundary diagnostics for logistic tails, fixed-epsilon crossover, and adaptive-method variants, delineating the scope of the proved smoothed-sign theory.

📄 PDF Abstract BibTeX arXiv:2607.06013

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

HR-NeuS: Recovering High-Frequency Surface Geometry via Neural Implicit Surfaces

2023-02-14 · Erich Liang, Kenan Deng, Xi Zhang, Chun-Kai Wang

Recent advances in neural implicit surfaces for multi-view 3D reconstruction primarily focus on improving large-scale surface reconstruction accuracy, but often produce over-smoothed geometries that lack fine surface det…

3D ReconstructionMulti-View 3D ReconstructionSurface ReconstructionVocal Bursts Intensity Prediction

A Mirror Descent Perspective of Smoothed Sign Descent

2024-10-18 · Shuyang Wang, Diego Klabjan

Recent work by Woodworth et al. (2020) shows that the optimization dynamics of gradient descent for overparameterized problems can be viewed as low-dimensional dual dynamics induced by a mirror map, explaining the implic…

The Implicit Length Bias of Label Smoothing on Beam Search Decoding

2022-05-02 · Bowen Liang, Pidong Wang, Yuan Cao

Label smoothing is ubiquitously applied in Neural Machine Translation (NMT) training. While label smoothing offers a desired regularization effect during model training, in this paper we demonstrate that it nevertheless …

Machine TranslationNMTTranslation

Direction Matters: On the Implicit Bias of Stochastic Gradient Descent with Moderate Learning Rate

2020-11-04 · ICLR 2021 1 · Jingfeng Wu, Difan Zou, Vladimir Braverman, Quanquan Gu

Understanding the algorithmic bias of \emph{stochastic gradient descent} (SGD) is one of the key challenges in modern machine learning and deep learning theory. Most of the existing works, however, focus on \emph{very sm…

Learning Theory

Characterizing the implicit bias via a primal-dual analysis

2019-06-11 · Ziwei Ji, Matus Telgarsky

This paper shows that the implicit bias of gradient descent on linearly separable data is exactly characterized by the optimal solution of a dual optimization problem given by a smoothed margin, even for general losses. …