Mostafa Milani

Mostafa Milani

Assistant Professor, Department of Computer Science, Western University

I am an Assistant Professor in the Department of Computer Science at Western University. I study how to make data systems more trustworthy and useful, helping people access, understand, and clean data despite noise, uncertainty, and privacy constraints.

My research connects quality and cleaning, privacy and fairness, provenance and explainability, and query understanding and recommendation. I develop principled methods and practical tools for these problems.

Before joining Western, I held postdoctoral positions at the University of British Columbia and McMaster University, and earned my Ph.D. from Carleton University.

Research interests

  • Quality and cleaning
  • Privacy and fairness
  • Provenance and explainability
  • Query understanding and recommendation
EmailGoogle ScholarDBLPLinkedInGitHub

News

  1. Welcome to Kiyana Jalali (MSc) and Melika Nasirian (PhD), who joined the group this year.

  2. Congratulations to Shadman Islam, Yingge Hu, Gautham Ramesh Babu, and Farhan Patel on completing their master’s degrees!

  3. Our survey and experimental analysis of heterogeneity in entity matching is published in Data & Knowledge Engineering.

    Read paper ↗

  4. Two ICDE workshop papers explore fairness-constrained private synthesis and human preferences in data cleaning.

    Structure Selection for Fairness-Constrained Differentially Private Data Synthesis ↗The Human Factor in Data Cleaning: Exploring Preferences and Biases ↗

  5. Our relation-stratified sampling paper for Shapley-value estimation received a Best Paper Award at an IEEE Big Data workshop.

    Read paper ↗

Selected publications

All publications →
DKE
Taxonomy of heterogeneity in entity matching.View figure ↗Effect of synonym injection on entity-matching performance across six dataset settings.View figure ↗

Heterogeneity in entity matching: A survey and experimental analysis

Mohammad Hossein Moslemi, Amir Mousavi, Behshid Behkamal, Mostafa Milani

Data Knowl. Eng. 164: 102575 (2026)

Paper ↗arXivPDF
BibTeX
@article{journals-dke-MoslemiMBM26,
  title = {Heterogeneity in entity matching: A survey and experimental analysis},
  author = {Mohammad Hossein Moslemi and Amir Mousavi and Behshid Behkamal and Mostafa Milani},
  year = {2026},
  journal = {Data & Knowledge Engineering},
  url = {https://doi.org/10.1016/j.datak.2026.102575},
  eprint = {2508.08076},
  archivePrefix = {arXiv},
  pdf = {https://arxiv.org/pdf/2508.08076},
  doi = {10.1016/j.datak.2026.102575}
}
EDBT
Query length and syntax-error classification outcomes for Llama3.View figure ↗Predicate count and missing-token classification outcomes for Gemini on SQLShare.View figure ↗

Evaluating SQL Understanding in Large Language Models

Ananya Rahaman, Anny Zheng, Mostafa Milani, Fei Chiang, Rachel Pottinger

EDBT 2025: 909-921

Paper ↗arXivPDFGitHub
BibTeX
@inproceedings{edbt-2025,
  title = {Evaluating SQL Understanding in Large Language Models},
  author = {Ananya Rahaman and Anny Zheng and Mostafa Milani and Fei Chiang and Rachel Pottinger},
  year = {2025},
  booktitle = {The International Conference on Extending Database Technology},
  url = {https://www.openproceedings.org/2025/conf/edbt/paper-211.pdf},
  eprint = {2410.10680},
  archivePrefix = {arXiv},
  pdf = {https://arxiv.org/pdf/2410.10680}
}
SIGMOD
A probabilistic transport plan linking original and repaired data tuples.View figure ↗OTClean performance under attribute noise, across four datasets.View figure ↗

OTClean: Data Cleaning for Conditional Independence Violations using Optimal Transport

Alireza Pirhadi, Mohammad Hossein Moslemi, Alexander Cloninger, Mostafa Milani, Babak Salimi

Proc. ACM Manag. Data 2(3): 160 (2024)

Paper ↗arXivPDF
BibTeX
@article{sigmod-2024,
  title = {OTClean: Data Cleaning for Conditional Independence Violations using Optimal Transport},
  author = {Alireza Pirhadi and Mohammad Hossein Moslemi and Alexander Cloninger and Mostafa Milani and Babak Salimi},
  year = {2024},
  journal = {Proceedings of the ACM on Management of Data},
  url = {https://dl.acm.org/doi/10.1145/3654963},
  eprint = {2403.02372},
  archivePrefix = {arXiv},
  pdf = {https://arxiv.org/pdf/2403.02372}
}
VLDB
Four scatter plots showing chase-termination checking runtime as the number of rules increases.View figure ↗

Semi-Oblivious Chase Termination for Linear Existential Rules: An Experimental Study

Marco Calautti, Mostafa Milani, Andreas Pieris

Proc. VLDB Endow. 16(11): 2858-2870 (2023)

Paper ↗arXivPDFGitHub
BibTeX
@article{vldb-2023,
  title = {Semi-Oblivious Chase Termination for Linear Existential Rules: An Experimental Study},
  author = {Marco Calautti and Mostafa Milani and Andreas Pieris},
  year = {2023},
  journal = {Proceedings of the VLDB Endowment},
  url = {https://dl.acm.org/doi/abs/10.14778/3611479.3611493},
  eprint = {2303.12851},
  archivePrefix = {arXiv},
  pdf = {https://arxiv.org/pdf/2303.12851}
}

Professional service

Read paper ↗