Research

Predicting and Interpreting Energy Barriers of Metallic Glasses with Graph Neural Networks

Haoyu Li*, Shichang Zhang*, Longwen Tang, Mathieu Bauchy, Yizhou Sun

Understanding the impact of their local structure on properties remains a challenge in the study of Metallic Glasses (MGs). In this study, we introduce a novel Symmetrized GNN (SymGNN) model, capable of predicting energy barriers while being invariant under transformations like rotations and reflections. SymGNNs aggregate information from these transformations, significantly improving energy barrier prediction accuracy and efficiency compared to traditional models.

[Paper] (ICML 2024)

Data Attribution for Diffusion Models: Timestep-induced Bias in Influence Estimation

Tong Xie*, Haoyu Li*, Andrew Bai, Cho-jui Hsieh

Data attribution methods help interpret how neural networks behave by linking the model behavior to their training data. We extend the first-order influence approximation, TracIn, to diffusion models by incorporating the denoising timestep dynamics. We demonstrate that this influence estimation may be biased due to dominating gradient norms. To this end, Diffusion-ReTrac with a renormalization technique is introduced, enabling notably more localized influence estimation and the targeted attribution of training samples.

[Paper] (TMLR 2024)