Efficient Sharpness-Aware Minimization for Molecular Graph Transformer Models

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) $\mathit{\text{gradient}}\mathit{\text{approximation}}$: we use the updating gradient of the previous step to approximate the perturbation gradient at the intermediate steps smoothly ($\mathbf{\text{increases}}\mathbf{\text{efficiency}}$); (ii) $\mathit{\text{loss}}\mathit{\text{landscape}}\mathit{\text{approximation}}$: we theoretically prove that the loss landscape of GraphSAM is limited to a small range centered on the expected loss of SAM ($\mathbf{\text{guarantees}}\mathbf{\text{generalization}}\mathbf{\text{performance}}$). 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