Simin Liu

Undergraduate Researcher in Physics & Artificial Intelligence
Nanjing University of Posts and Telecommunications

My research interests lie at the intersection of artificial intelligence, physics, materials science, and biomedical engineering. I am particularly interested in applying machine learning and computational methods to interdisciplinary scientific and engineering problems.

Email GitHub CV
Simin Liu

About Me

I am an undergraduate student in Physics at Nanjing University of Posts and Telecommunications (NUPT), expected to graduate in 2027.

My research has primarily focused on machine-learning-assisted superconducting materials discovery. I have worked on interpretable machine learning, graph neural networks, and materials informatics for superconducting critical temperature prediction.

I am also exploring interdisciplinary research in biomedical signal processing, intelligent sensing, and neuromorphic devices. I enjoy learning new techniques and applying computational approaches to problems across different scientific fields.

Machine Learning Graph Neural Networks AI for Science Materials Informatics Biomedical Signal Processing Intelligent Biosensing

Research Experience

AI-Assisted Discovery of Superconducting Materials

Nanjing University of Posts and Telecommunications · Advisor: Prof. Bin Li · Jun. 2024 – Present
  • Developing machine-learning models to predict superconducting critical temperature (Tc) and investigate the relationships among material composition, crystal structures, and superconducting properties.
  • Constructed more than 300 physical descriptors from material composition and CIF crystal structures, and compared six regression models including Random Forest, XGBoost, and Gradient Boosting.
  • Applied SHAP to investigate important physical factors affecting superconducting critical temperature. The XGBoost model achieved a test R2 of 0.745.
  • Participated in developing HGTC-Net, a graph neural network incorporating superconducting category priors. Semantic gating improved the test R2 from 0.7448 for direct graph regression to 0.8809.
  • This research has resulted in two first/co-first-author papers on machine learning for superconducting materials.

Rapid Calibration for Cuffless Blood Pressure Monitoring

Southeast University · Advisor: Prof. Kuo Yang · Apr. 2026 – Present
  • Investigating rapid calibration and drift compensation methods for long-term cuffless blood pressure monitoring using multimodal physiological signals.
  • Processed physiological signals including photoplethysmography (PPG) and heart rate, and developed a preliminary pipeline covering signal preprocessing, feature extraction, and blood pressure prediction.
  • Exploring personalized calibration using multimodal physiological information and adaptive calibration methods for signal drift during long-term monitoring.

Self-Powered Bidirectional Synaptic Device Based on Cs-Doped Perovskite/P3HT Heterojunction

Nanjing University of Posts and Telecommunications · Advisor: Prof. Wen Huang · Apr. 2026 – Jun. 2026
  • Investigated optoelectronic response and synaptic plasticity in Cs-doped perovskite/P3HT heterojunction devices.
  • Conducted experimental data analysis and visualization under different optical powers, pulse widths, frequencies, and pulse numbers.
  • Observed IPSC peaks of approximately 4500 nA and stable device responses across 50 repeated cycles.
  • Contributed to manuscript writing, revision, scientific visualization, and figure preparation.

Literature Study on Superconducting Hydrides

Nanjing University of Posts and Telecommunications · Advisor: Prof. Bin Li · Jul. 2025 – Dec. 2025
  • Reviewed recent progress in high-temperature superconducting hydrides, with particular emphasis on ternary hydrides and strategies for extending superconductivity toward lower pressures and ambient conditions.
  • Participated in literature review, data collection, analysis, and manuscript preparation for a review published in Annalen der Physik.

Publications

Understanding Tc Variation in Isostructural Hydrides: Interpretable Machine Learning with Physical Descriptors
Simin Liu, J. Wang, B. Li, J. Zhai, M. Wu, Z. Cui, Y. Ren, Y. Zhang, S. Liu
Physica B: Condensed Matter, 737, 418748 (2026). Published
Hierarchical Graph Learning with Superconducting Category Priors for Tc Prediction
T. Zhou, Simin Liu, B. Li, S. Xu, M. Wu, Z. Cui
Physical Review B, 2026. Accepted
Co-first author · Accepted Aug. 28, 2026
Design and Discovery of High-Temperature Superconducting Ternary Hydrides: From High Pressure to Ambient Conditions
B. Li, J. Zhai, Z. Cao, Simin Liu, et al.
Annalen der Physik, 538, e00462 (2026). Published

Education

Nanjing University of Posts and Telecommunications

B.Sc. in Physics
Sep. 2023 – Jun. 2027 (Expected)

Selected Honors & Awards

Skills

AI & Machine Learning
Machine Learning, Deep Learning, Graph Neural Networks, XGBoost, SHAP, PyTorch, Scikit-learn
Materials Informatics
Matminer, Materials Descriptors, Feature Engineering, VASP
Programming
Python, C, basic Java, basic C++
Research Tools
Linux, Git, LaTeX, Origin, Adobe Illustrator