Papers › LipschitzLR: Using theoretically computed adaptive learning rates for fast convergence

LipschitzLR: Using theoretically computed adaptive learning rates for fast convergence

20 Feb 2019arXiv:1902.07399archive 2025-07-28

Rahul Yedida, Snehanshu Saha, Tejas Prashanth

Optimizing deep neural networks is largely thought to be an empirical process, requiring manual tuning of several hyper-parameters, such as learning rate, weight decay, and dropout rate. Arguably, the learning rate is the most important of these to tune, and this has gained more attention in recent works. In this paper, we propose a novel method to compute the learning rate for training deep neural networks with stochastic gradient descent. We first derive a theoretical framework to compute learning rates dynamically based on the Lipschitz constant of the loss function. We then extend this framework to other commonly used optimization algorithms, such as gradient descent with momentum and Adam. We run an extensive set of experiments that demonstrate the efficacy of our approach on popular architectures and datasets, and show that commonly used learning rates are an order of magnitude smaller than the ideal value.

PaperPDFCode

Code

yrahul3910/adaptive-lr-dnn officialmentioned in paper report
sahamath/SYM-Net mentioned on GitHub report
sahamath/sym-netv1 mentioned on GitHub report
tej-prash/GeneExpression mentioned on GitHub report
yrahul3910/symnet mentioned on GitHub report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Handwritten Digit RecognitionObject Detection

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AdamDropout

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections