Milad Hoseinpour

EECS Ph.D. Candidate at the University of Michigan.

prof_pic.jpg

About Me

I am a Ph.D. candidate in the Department of Electrical Engineering and Computer Science at the University of Michigan, fortunate to be advised by Prof. Vladimir Dvorkin. I received my M.Sc. and B.Sc. degrees in Electrical Engineering from Tarbiat Modares University and Babol Noshirvani University of Technology, respectively.

Research Interests

My primary research interests lie in Generative Models, Trustworthy ML, and Information Theory. I focus on developing reliable data-driven decision-making methods with rigorous guarantees on uncertainty, robustness, and privacy. Methodologically, I use ideas from diffusion models, differential privacy, conformal prediction, and sequential inference. I am particularly interested in bridging foundational theory with practical challenges in modern power and energy systems.

Feel free to email me if you come across any research interests we may have in common. I am always open to potential collaborations.

news

Jan 27, 2026 Our paper DiffOPF: Diffusion Solver for Optimal Power Flow has been accepted to the Power Systems Computation Conference (PSCC) 2026.
Jan 20, 2026 Our new paper Outage Identification from Electricity Market Data: Quickest Change Detection Approach has been released on arXiv!
Dec 10, 2025 I am honored to have been selected as a 2026 Fellow in the Institute for Energy Solutions’ PhD and Postdoctoral Fellowship Program.
Oct 27, 2025 I will be presenting at the 2025 INFORMS Annual Meeting in Atlanta: Synthesizing Power Flow Datasets via Constrained Diffusion Models (slides, paper).
Oct 15, 2025 Our new paper DiffOPF: Diffusion Solver for Optimal Power Flow has been released on arXiv!
Oct 02, 2025 I am serving as an Organizer for the 2026 Michigan Student Symposium for Interdisciplinary Statistical Sciences (MSSISS).
Aug 25, 2025 Our new paper Constrained Diffusion Models for Synthesizing Representative Power Flow Datasets has been released on arXiv!
Aug 21, 2025 I began serving as a Leading Committee Member for Academic Careers in ECE (ACE), a group within the EECS Department at the University of Michigan dedicated to supporting Ph.D. students and post-doctoral researchers interested in pursuing future faculty careers.
Jun 12, 2025 Our paper Domain-constrained diffusion models to synthesize tabular data: A case study in power systems is accepted for presentation at the 2025 ICML workshop DataWorld: Unifying Data Curation Frameworks Across Domains (poster).
May 21, 2025 I was honored to receive the Student Poster Competition Award at the Institute for Energy Solutions2025 Energy Symposium.

selected publications

  1. arXiv
    DiffOPF: Diffusion Solver for Optimal Power Flow
    Milad Hoseinpour and Vladimir Dvorkin
    arXiv preprint arXiv:2510.14075, 2025
  2. arXiv
    Constrained Diffusion Models for Synthesizing Representative Power Flow Datasets
    Milad Hoseinpour and Vladimir Dvorkin
    2025
  3. ICML
    Domain-Constrained Diffusion Models to Synthesize Tabular Data: A Case Study in Power Systems
    Milad Hoseinpour and Vladimir Dvorkin
    In ICML 2025 Workshop on DataWorld: Unifying Data Curation Frameworks Across Domains, 2025
  4. EPSR
    Prosumers’ Cost Recovery in Peer-to-Peer Electricity Markets
    Milad Hoseinpour and Mahmoud-Reza Haghifam
    Electric Power Systems Research, 2024
  5. TSG
    Privacy-Preserving and Approximately Truthful Local Electricity Markets: A Differentially Private VCG Mechanism
    Milad Hoseinpour, Mohammad Hoseinpour, Mahdi Haghifam, and 1 more author
    IEEE Transactions on Smart Grid, 2023