LipschitzLR: Using theoretically computed adaptive learning rates for fast convergence
This addresses the need for automated hyperparameter tuning in deep learning, offering a theoretical approach to improve convergence speed, though it appears incremental as it builds on existing optimization algorithms.
The paper tackles the problem of manually tuning learning rates in deep neural network optimization by proposing a method to compute adaptive learning rates based on the Lipschitz constant of the loss function, showing that commonly used rates are an order of magnitude smaller than ideal values.
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.