Computational Physicist, Data Scientist
Computational Physicist, Data Scientist

Biography

Gihan Panapitiya specializes in computational materials design and discovery, leveraging first-principles theory calculations and deep learning. Since joining Pacific Northwest National Laboratory in late 2019, his research has centered on developing deep learning models for battery material discovery, predicting cancer drug response, and creating autonomous experimentation techniques based on generative artificial intelligence

Research Interest

  • Computational Condensed Matter Physics
  • Density Functional Theory
  • Deep Learning
  • Materials Discovery
  • Graph Neural Networks
  • Transformers

Education

  • PhD in physics, West Virginia University
  • MS in physics, University of Akron
  • BS in computational physics, University of Colombo

Publications

2025

  • Kong J., G.U. Panapitiya, and E.G. Saldanha. 2025. "Extracting Material Property Measurements from Scientific Literature With Limited Annotations." Journal of Chemical Information and Modeling 65, no. 10:4906-4917. PNNL-SA-194296. doi:10.1021/acs.jcim.4c01352

2023

  • Yin T., G.U. Panapitiya, E.D. Coda, and E.G. Saldanha. 2023. "Evaluating Uncertainty-Based Active Learning for Accelerating the Generalization of Molecular Property Prediction." Journal of Cheminformatics 15. PNNL-SA-179045. doi:10.1186/s13321-023-00753-5

2022

  • Gao P., A. Andersen, J.P. Sepulveda, G.U. Panapitiya, A.M. Hollas, E.G. Saldanha, and V. Murugesan, et al. 2022. "SOMAS: a platform for data-driven material discovery in redox flow battery development." Scientific Data 9. PNNL-SA-161978. doi:10.1038/s41597-022-01814-4
  • Panapitiya G.U., M.K. Girard, A.M. Hollas, J.P. Sepulveda, V. Murugesan, W. Wang, and E.G. Saldanha. 2022. "Evaluation of Deep Learning Architectures for Aqueous Solubility Prediction." ACS Omega 7, no. 18:15695-15710. PNNL-SA-161618. doi:10.1021/acsomega.2c00642

2021

  • Panapitiya G.U., F.C. Parks, J.P. Sepulveda, and E.G. Saldanha. 2021. "Extracting Material Property Measurement Data from Scientific Articles." In Proceedings of the 2021 Conference on Empirical Methods i Natural Language Processing (EMNLP 2021), November 7-11, 2021, Online and Punta Cana, Dominican Republic, 5393-5402. Stroudsburg, Pennsylvania:Association for Computational Linguistics. PNNL-SA-162563. doi:10.18653/v1/2021.emnlp-main.438