Efficient Sharpness-Aware Minimization for Molecular Graph Transformer Models

TL;DR

This paper introduces GraphSAM, an efficient algorithm for molecular graph transformer training that reduces computational costs of sharpness-aware minimization while maintaining or improving generalization performance.

Contribution

The paper proposes GraphSAM, a novel method that approximates SAM's perturbation gradient using previous gradients and proves its loss landscape is tightly bounded, enhancing efficiency and generalization.

Findings

GraphSAM reduces training time compared to traditional SAM.

It improves generalization performance on multiple datasets.

Theoretical analysis guarantees bounded loss landscape.

Abstract

Sharpness-aware minimization (SAM) has received increasing attention in computer vision since it can effectively eliminate the sharp local minima from the training trajectory and mitigate generalization degradation. However, SAM requires two sequential gradient computations during the optimization of each step: one to obtain the perturbation gradient and the other to obtain the updating gradient. Compared with the base optimizer (e.g., Adam), SAM doubles the time overhead due to the additional perturbation gradient. By dissecting the theory of SAM and observing the training gradient of the molecular graph transformer, we propose a new algorithm named GraphSAM, which reduces the training cost of SAM and improves the generalization performance of graph transformer models. There are two key factors that contribute to this result: (i) \textit{gradient approximation}: we use the updating gradient of the previous step to approximate the perturbation gradient at the intermediate steps smoothly (\textbf{increases efficiency}); (ii) \textit{loss landscape approximation}: we theoretically prove that the loss landscape of GraphSAM is limited to a small range centered on the expected loss of SAM (\textbf{guarantees generalization performance}). The extensive experiments on six datasets with different tasks demonstrate the superiority of GraphSAM, especially in optimizing the model update process. The code is in:https://github.com/YL-wang/GraphSAM/tree/graphsam