Acceleration with Inexact Gradients: Convergence Rates and Error Accumulation
发布时间：2026年10月21日
浏览次数：15
发布者: He Liu
主讲人： 刘茵（数学中心）
活动时间： 从 2026-10-23 15:00 到 16:00
场地： 北京国际数学研究中心，全斋全29教室
Accelerated gradient methods achieve faster convergence rates than standard gradient descent in smooth convex optimization, but they can be more sensitive to gradient estimation errors. This raises a natural question: how much acceleration can be preserved when only approximate gradients are available?
In this talk, I will explain the mechanism of error accumulation under an absolute error condition. Using the Performance Estimation Problem (PEP) framework, we derive nonasymptotic convergence bounds that quantify the overall impact of these errors. Our results reveal a trade-off between acceleration and error accumulation, providing insights into the choice of algorithmic parameters.
