Ya Ju Fan is a computational scientist in the Center for Applied Scientific Computing at Lawrence Livermore National Laboratory. Her research focuses on developing data-driven mathematical models and statistical methods for scientific data analysis. She is particularly interested in understanding the mathematical foundations of machine learning and harnessing these methods to solve practical scientific problems. Her areas of expertise include uncertainty quantification, anomaly detection, autoencoders, graph neural networks, and generative models.

Fan collaborates with interdisciplinary teams to address complex problems across a wide range of scientific domains. She develops machine learning solutions grounded in careful scientific judgment and tailored to the needs of each application. Her work has supported research in power systems, cancer genomics, drug discovery, interatomic potentials, environmental analysis of the Marshall Islands, wind energy generation, additive manufacturing, and EEG time-series analysis.

A central theme of Fan’s research is advancing machine learning models by addressing their limitations and identifying new opportunities for improvement. She developed the Autoencoder Node Saliency method to interpret the behavior of individual neurons in neural networks. She is also a co-inventor of Support Feature Machines, a class of optimization models for multivariate time-series anomaly detection. In generative AI, she has incorporated constraints and latent-space predictive capabilities to guide the generation of drug compounds with desired properties. She contributed to the development of attention-based graph encoder methods for learning interatomic potentials.

Fan has published her research in leading journals and conferences, including Pattern Recognition, Operations Research, INFORMS Journal on Computing, SIAM Conference on Optimization, and the ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD).

Ph.D. in Industrial and Systems Engineering, Rutgers University, New Jersey

M.Sc. in Decision Science/Operations Research, University of Wisconsin-Madison

Y. J. Fan, J. E. Allen, K. S. McLoughlin, D. Shi, B. J. Bennion, X. Zhang, F. C. Lightstone. "Evaluating point-prediction uncertainties in neural networks for protein-ligand binding prediction". Artificial Intelligence Chemistry, Volume 1, Issue 1, 2023. https://doi.org/10.1016/j.aichem.2023.100004.

Y. J. Fan, J. Allen, S. A. Jacobs and B. Van Essen. “Distinguishing between Normal and Cancer Cells Using Autoencoder Node Saliency”. Second ISC HPC Applications in Precision Medicine Workshop. June 28, 2018.

Y. J. Fan. “Autoencoder Node Saliency: Selecting Relevant Latent Representations”. Pattern Recognition, Volume 88, pp.643-653. April 2019. https://doi.org/10.1016/j.patcog.2018.12.015

O. Seref, Y. J. Fan, E. Borenstein, W. A. Chaovalitwongse. “Information-theoretic feature selection with discrete k-median clustering”. Annals of Operations Research. 263(1-2): 93-118, 2018.

Y. J. Fan and C. Kamath. "A comparison of compressed sensing and sparse recovery algorithms applied to simulation data". Statistics, Optimization, and Information Computing, Vol. 4, Issue 3, pp. 194-213. September 2016. 

Y. J. Fan and C. Kamath. "Detecting ramp events in wind energy generation using affinity evaluation on weather data". Statistical Analysis and Data Mining, Volume 9, issue 3, pp. 155–173. June 2016.

Y. J. Fan and C. Kamath. "Practical Considerations in Applying Compressed Sensing to Simulation Data". Data Compression Conference (DCC). April 2015.

Y. J. Fan and C. Kamath. “Identifying and Exploiting Diurnal Motifs in Wind Generation Time Series Data”. International Journal of Pattern Recognition and Artificial Intelligence , Vol 29, Number 2, pp. 1550012-1 - 1550012-25. March 2015.

C. Kamath and Y. J. Fan. "Incremental SVD for Insight into Wind Generation". 13th International Conference on Machine Learning and Applications ICMLA 2014. December 2014.

 O. Seref, Y. J. Fan, W. A. Chaovalitwongse. ”Mathematical Programming Formulations and Algorithms for Discrete k-Median Clustering of Time-Series Data”. INFORMS Journal on Computing. 26(1): 160-172, 2014.

C. Kamath and Y. J. Fan. “Finding Motifs in Wind Generation Time Series Data”. 11th International Conference on Machine Learning and Applications ICMLA 2012. December 2012.

C. Kamath and Y. J. Fan. “Using Data Mining Techniques to Enable Integration of Wind Energy on the Power Grid”.  Statistical Analysis & Data Mining. Volume 5, Issue 5, pp 410-427, October 2012.

Y.  J. Fan and C. Kamath. “On the Selection of Dimension Reduction Techniques for Scientific Applications”.  Annals of Information Systems. To Appear. August 2012.

Y. J. Fan, and W. A. Chaovalitwongse, "Optimizing Feature Selection to Improve Medical Diagnosis". Annals of Operations Research on Data Mining, 174(1): 169-183, 2010.

W. A. Chaovalitwongse, Y. J. Fan, and R. C. Sachdeo. "Novel Optimization Models for Abnormal Brain Activity Classification". Operations Research, 56(6): 1450-1460, December, 2008.

W. A. Chaovalitwongse, Y. J. Fan, and R. C. Sachdeo. "On the Time Series K-Nearest Neighbor Classification of Abnormal Brain". IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans, 37(6): 1005-1016, November, 2007.

W. A. Chaovalitwongse, Y. J. Fan, and R. C. Sachdeo. Support Feature Machine for Classification of Abnormal Brain Activity. The Thirteenth ACM SIGKDD International Conference On Knowledge Discovery and Data Mining (SIGKDD 2007), pp. 113-122.

Presentations

“Distinguishing between Normal and Cancer Cells Using Autoencoder Node Saliency”. Second ISC HPC Applications in Precision Medicine Workshop. Frankfurt, Germany. June 28, 2018

"Incremental SVD for Insight into Wind Generation", 13th International Conference on Machine Learning and Applications, Detroit, MI. December 3-5, 2014.

“Exploiting Motifs and Anomalies in Streaming Data”, Department of Energy Applied Mathematics Program Meeting, Albuquerque, NM. August 5-8, 2013.

“Detecting Changes in Weather Data Streams for Wind Energy Prediction”, SIAM Annual Meeting, San Diego, CA. July 8-12, 2013.

“Determining Number of Motifs in Wind Generation Time Series Data”, SIAM Annual Meeting, San Diego, CA. July 8-12, 2013.

“Finding Motifs in Wind Generation Time Series Data”, 11th International Conference on Machine Learning and Applications, Boca Raton, FL. December 12-15, 2012.

 “A Heuristic for the Local Region Covering Problem”, 21st International Symposium on Mathematical Programming, Berlin, Germany. August 19-24, 2012.

“A Comparison of Dimensionality Reduction Techniques in Scientific Applications”, Center for Advanced Signal and Image Sciences, Livermore, CA. May 23, 2012.

 “A Comparison of Dimensionality Reduction Techniques in Scientific Applications”, SIAM Conference on Uncertainty Quantification, Raleigh, NC. April 2-5, 2012.

 “Intrinsic Dimensionality Using Non-linear Dimension Reduction Techniques”, Institute for Operations Research and Management Sciences Annual Meeting, Charlotte, NC. November 13-16, 2011.

“A Comparison of Non-linear Techniques for Dimension Reduction”, LLNL Postdoc Poster Symposium. June 1, 2011.

“Medical Data Classification Via Optimizing Feature Selection”, Institute for Operations Research and the Management Sciences (INFORMS) Annual Meeting, Washington, D.C. October 2008.

“Classification of Normal and Abnormal EEG Signals Using K-Nearest Neighbor Rule in Support Feature Machine”, Conference on Computational Neuroscience, University of Florida, Gainesville, FL. February 2008.

“On the Time Series K-Nearest Neighbor Classification of Abnormal Brain Activity”, Institute for Operations Research and the Management Sciences (INFORMS) Annual Meeting, Pittsburgh, PA. November 2006.

• Pierskalla Best Paper Award, INFORMS Annual Meeting, Washington, D.C., October 2008.
• Transportation Coordinating Council / Federal Transit Administration (TCC/FTA) Fellowship, Fall 2007 and Spring 2008
• Kuhl Memorial Engineering Fellowship, Rutgers Graduate School, Fall 2006