OCLGNAMLOct 10, 2020

AEGD: Adaptive Gradient Descent with Energy

arXiv:2010.05109v213 citations
Originality Incremental advance
AI Analysis

This work addresses optimization challenges in machine learning, particularly for non-convex problems like deep neural networks, by introducing a robust and tunable method, though it appears incremental as it builds on existing gradient descent frameworks.

The authors tackled the problem of optimizing non-convex objective functions by proposing AEGD, a new gradient-based algorithm that dynamically updates an energy variable, resulting in unconditional energy stability and proven convergence rates for both non-convex and convex settings, with experimental results showing comparable or better generalization performance than SGD with momentum for deep neural networks.

We propose AEGD, a new algorithm for first-order gradient-based optimization of non-convex objective functions, based on a dynamically updated energy variable. The method is shown to be unconditionally energy stable, irrespective of the step size. We prove energy-dependent convergence rates of AEGD for both non-convex and convex objectives, which for a suitably small step size recovers desired convergence rates for the batch gradient descent. We also provide an energy-dependent bound on the stationary convergence of AEGD in the stochastic non-convex setting. The method is straightforward to implement and requires little tuning of hyper-parameters. Experimental results demonstrate that AEGD works well for a large variety of optimization problems: it is robust with respect to initial data, capable of making rapid initial progress. The stochastic AEGD shows comparable and often better generalization performance than SGD with momentum for deep neural networks.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes