Publications

See full publications in Google Scholar.

SimXRD-4M: Big Simulated X-ray Diffraction Data and Crystalline Symmetry Classification Benchmark

Published in ICLR, 2025

In this paper, we developed the largest open-source simulated X-ray diffraction database (SimXRD). SimXRD comprises 4,065,346 simulated powder XRD patterns, representing 119,569 unique crystal structures under 33 simulated conditions that reflect real-world variations. We benchmark 21 sequence models in both in-library and out-of-library scenarios and analyze the impact of class imbalance in long-tailed crystal label distributions. Remarkably, we find that: (1) current neural networks struggle with classifying low-frequency crystals, particularly in out-of-library situations; (2) models trained on SimXRD can generalize to real experimental data.

Recommended citation: Cao, bin and Liu, Yang and Zheng, Zinan and Tan, Ruifeng and Li, Jia and Zhang, Tong-Yi. "SimXRD-4M: Big Simulated X-ray Diffraction Data and Crystalline Symmetry Classification Benchmark", ICLR 2025.
Download Paper

Forecasting battery degradation trajectory under domain shift with domain generalization

Published in Energy Storage Materials, 2024

In this paper, we conceive that differences in battery operating conditions can be treated as domain shift, and pioneer domain generalization for battery degradation trajectory forecasting. Specifically, meta-learning-based and task-driven domain generalization techniques are utilized to attack the domain shift. The effectiveness of the proposed method is demonstrated on 203 cells of various operating conditions and chemistries. Our work also spotlights the potential interplay between artificial intelligence and domain knowledge.

Recommended citation: Tan, R., Lu, X., Cheng, M., Li, J., Huang, J., & Zhang, T. Y. (2024). Forecasting battery degradation trajectory under domain shift with domain generalization. Energy Storage Materials, 72, 103725.
Download Paper

Unravelling Thermal and Enthalpy Evolutions of Commercial Sodium-Ion Cells upon Cycling Ageing via Fiber Optic Sensors

Published in Journal of The Electrochemical Society, 2023

In this paper, we demonstrate the feasibility of using fiber Bragg grating sensors to operando monitor the thermal (temperature and heat) evolutions of commercial 18650 sodium-ion cells during long-term cycling ageing. With the delicate heat deconvolution, the evolutions of entropy and overpotential heat rates upon the cycling ageing are decoded, while the ageing-driven changes in overpotential components are further analysed. Drawing also on thermodynamics, high-resolution enthalpy profiles are computed from operando heat and voltage profiles, enabling to track and unravel redox variations during the cycling ageing.

Recommended citation: Huang, J., Delacourt, C., Desai, P., Gervillié-Mouravieff, C., Blanquer, L. A., Tan, R., & Tarascon, J. M. (2023). Unravelling Thermal and Enthalpy Evolutions of Commercial Sodium-Ion Cells upon Cycling Ageing via Fiber Optic Sensors. Journal of The Electrochemical Society, 170(9), 090510.
Download Paper

PMGNN: A Pioneer-Master Graph Neural Network for Graph Classification

Published in IJCNN, 2022

This paper proposes a GNN that can consider multi-hop information for graph classification task. The effectiveness of the proposed method is demonstrated on four real-world datasets.

Recommended citation: Tan, Ruifeng, and Yuanyuan Zhu. "PMGNN: A Pioneer-Master Graph Neural Network for Graph Classification." 2022 International Joint Conference on Neural Networks (IJCNN). IEEE, 2022.
Download Paper