1 of 6

Data Science Lab at SDSU

Director: Hajar Homayouni

hhomayouni@sdsu.edu

2 of 6

Objectives

Improving the quality and reliability of medical data, which is essential for informed decision-making.

Enhancing the interpretability of machine learning models to gain insights into their decision-making processes.

Leading a diverse team of students from various academic levels.

3 of 6

Data Science Lab

Synthetic Medical Data Generation

Anomaly Detection and Explanation

EDA of Medical Research Publications

4 of 6

Privacy-Preserving Synthetic Medical Data Generation

  • A. Kurakova, H. Homayouni. “A Comprehensive Evaluation Framework for Synthetic Medical Tabular Data Generation”, Journal of Biomedical Informatics JBI: Special issue Data Generation in Healthcare Environments, 2025.
  • Ilaty, H. Shirazi, H. Homayouni. “SynLLM: A Comparative Analysis of Large Language Models for Medical Tabular Synthetic Data Generation”, Neural Information Processing Systems Workshop on GenAI for Health (NeurIPS-GenAI4Health), 2025.
  • A. Jnaini, H. Shirazi, H. Homayouni. “Synergy of GPT-3 Summarization and Vision-Encoder-Decoder for Chest X-ray Captioning”, in the IEEE Canadian Conference On Electrical and Computer Engineering (IEEE CCECE), 2024.
  • H. Homayouni, M. Pourebadi, H. Shirazi. ”Poster: Federated Multimodal Medical Data Generation”, presented in the Network and Distributed System Security Symposium (NDSS), 2024.
  • S. Kaur, S. Kumar, H. Homayouni. “High-Resolution COVID-19 X-Ray Generator”, Health Informatics and Knowledge Management Conference- HIKM, pp. 151-159, 2023.

ALM: Augmented Lagrangian Method

5 of 6

Anomaly Detection and Explanation

H. Homayouni, S. K. Polu, H. Shirazi. “CACL: Context-Aware Contrastive Learning for Semantic Modeling of Anomalies in EHR Data,” in the 17th ACM Conference on Bioinformatics, Computational Biology, and Health Informatics (ACM-BCB), pp. 1—10, 2026, doi: 10.1145/3807503.3819465.

H. Homayouni, H. Aghayarzadeh, I. Ray, and H. Shirazi, “Anomaly Detection and Interpretation from Tabular Data Using Transformer Architecture,” in the International Conference on Data Mining Workshop on Data-Centric AI (ICDM-DCAI), 2024, doi: 10.1109/ICDMW65004.2024.00091.

H. Homayouni, S. Sekeh, H. Shirazi. “GrEAt: Generalizable and Noise-Robust Energy-Based Anomaly Detection for Healthcare and Biomedical Tabular Data,” ACM Transactions on Computing for Healthcare, submitted.

6 of 6

EDA of Medical Publications

H. Shirazi, S. Bharath, E. Tasdighi, H. Homayouni, “From Atlas to Assistant: Mapping Two Decades of Cardiovascular Research and Enabling Reliable AI-Powered Knowledge Access,” Health Information Science and Systems, Special Issue: Toward Trustworthy and Generalizable AI in Healthcare: Advances in Explainability, Causal Learning and Cost-effectiveness Artificial Intelligence (AI), revised and resubmitted.

R. Talukder, H. Shirazi, A. Shishodia, S. Pandit, Hajar Homayouni, “Transformer-Based Topic Mapping and GPT-Driven Hierarchical Taxonomy in Cardiovascular Research,” in the IEEE International Conference on Data Mining workshop on AI-ready Data for Science Discovery (ICDM-ADSD), 2025, doi: 10.1109/ICDMW69685.2025.00019

D. Rozenshteyn , H. Homayouni. “Gender Gap Analysis in Cardiovascular Research”, IEEE Healthcom, 2024.

E. Navarro, H. Homayouni. “Topic Modeling in Cardiovascular Research Publications”, in the Undergraduate Consortium at the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD-UC), 2023.