Soroush Saghafian is an Iranian-American operations researcher and an associate professor of Public Policy at the John F. Kennedy School of Government at Harvard University.[1] He is best known for developing and applying artificial intelligence and operations research methods to improve healthcare systems.[2][3]

Soroush Saghafian
Born
NationalityIranian-American
Alma materUniversity of Michigan (Ph.D.)
University of Michigan (M.S.)
Sharif University of Technology (M.S.)
Known forAmbiguous Partially Observable Markov Decision Processes
Ambiguous Dynamic Treatment Regimes
AwardsINFORMS MSOM Young Scholar Prize (2021)
INFORMS Mehrotra Research Excellence Award (2020)
Scientific career
FieldsOperations Research

Management Science

Health Policy
InstitutionsHarvard University

Education

edit

Saghafian obtained an M.S. degree in Industrial Engineering from Sharif University of Technology in 2005 and an M.S. in Mathematics from the University of Michigan in 2009. He holds a Ph.D. in Industrial and Operations Engineering from the University of Michigan, awarded in 2012.[4]

Career

edit

Saghafian is an associate professor of Public Policy at the John F. Kennedy School of Government at Harvard University.[5] He is the founder of the Public Impact Analytics Science Lab (PIAS-Lab), dedicated to applying data science to complex societal issues.[6] His work is noted for its practical impact, influencing both healthcare policy and clinical practices. In 2024, PIAS-Lab received a $3 million grant from the U.S. Department of Defense to develop AI-driven personalized treatments for melanoma in collaboration with the Dana-Farber Cancer Institute.[7][8]

He serves on the editorial boards of several top-tier journals, including Management Science[9] and Operations Research.[10]

Notable Contributions and Research Impact

edit

Saghafian's research bridges advanced AI methods with real-world applications, notably within the healthcare sector. Saghafian developed the concepts of "Ambiguous Partially Observable Markov Decision Processes (APOMDP)" and "Ambiguous Dynamic Treatment Regimes" in operations research, methods that have improved clinical decision-making under uncertain conditions.[11][12] This work, funded by the National Science Foundation, resulted in the various new findings for treating patients who undergo transplantation and develop risks of New-Onset Diabetes After Transplantation (NODAT).[13]

His FDA-related research applies AI to improve medical device safety, significantly reducing recall rates by integrating human insights with machine learning algorithms. This approach could yield billions in healthcare cost savings by streamlining FDA review processes.[14]

Media Features

edit

Saghafian has been featured in various media outlets for his work and expertise in healthcare policy and AI in medicine. In an interview with Fast Company, he discussed the transformative potential of AI in healthcare.[15] Euronews highlighted his expertise on the use of AI to predict patient responses to antidepressant treatment.[16]

He has also provided expert analysis on healthcare policy and AI issues in PBS NewsHour[17], NBC News[18], WJCL (TV)[19], and WFMJ.[20]

Selected Publications

edit
  • Saghafian, S. (2018). "Ambiguous partially observable Markov decision processes: Structural results and applications." Journal of Economic Theory, 178, 1–35. doi:10.1016/j.jet.2018.08.006.
  • Saghafian, S. (2023). "Ambiguous Dynamic Treatment Regimes: A Reinforcement Learning Approach." Management Science. doi:10.1287/mnsc.2022.00883.
edit

References

edit
  1. ^ "Soroush Saghafian, Associate Professor". Harvard Kennedy School. 2 May 2024. Retrieved 26 September 2024.
  2. ^ "Using AI for Public Impact: Insights from Dr. Soroush Saghafian". Fast Company. 2024. Retrieved 26 September 2024.
  3. ^ "AI May Help to Predict How Patients Respond to Antidepressant Treatment". Euronews. 12 February 2024. Retrieved 26 September 2024.
  4. ^ "Soroush Saghafian CV". Harvard Scholar. Retrieved 26 September 2024.
  5. ^ "Soroush Saghafian, Associate Professor". Harvard Kennedy School. 2 May 2024. Retrieved 26 September 2024.
  6. ^ "Public Impact Analytics Science Lab". Harvard Scholar. Retrieved 26 September 2024.
  7. ^ "Public Impact Analytics Science Lab at the Harvard Kennedy School Receives $3 Million From DOD". The Harvard Crimson. 5 September 2024. Retrieved 26 September 2024.
  8. ^ "FY23 Team Science Award". Congressionally Directed Medical Research Programs. Retrieved 26 September 2024.
  9. ^ "Management Science Editorial Board". INFORMS. Retrieved 26 September 2024.
  10. ^ "Operations Research Editorial Board". INFORMS. Retrieved 26 September 2024.
  11. ^ Saghafian, Soroush (2018). "Ambiguous partially observable Markov decision processes: Structural results and applications". Journal of Economic Theory. 178: 1–35. doi:10.1016/j.jet.2018.08.006.
  12. ^ Saghafian, Soroush (2023). "Ambiguous Dynamic Treatment Regimes: A Reinforcement Learning Approach". Management Science. arXiv:2112.04571. doi:10.1287/mnsc.2022.00883.
  13. ^ "Data-Driven Management of Post-Transplant Medications". National Science Foundation.
  14. ^ "Transforming FDA Clearance: How AI and Human Insight Can Improve Medical Device Safety". Devdiscourse. 18 July 2024. Retrieved 26 September 2024.
  15. ^ "Five Takeaways from an AI Pioneer About Its Potential Impact in Healthcare". Fast Company. 2024. Retrieved 26 September 2024.
  16. ^ "AI May Help to Predict How Patients Respond to Antidepressant Treatment". Euronews. 12 February 2024. Retrieved 26 September 2024.
  17. ^ "Decades after Historic Black Hospital Closes, Former Nurses Fight to Keep the Memory Alive". PBS NewsHour. 15 December 2023. Retrieved 26 September 2024.
  18. ^ "Atlanta's Health Care System Is Strained by Major Hospital's Closing, Doctors and Patients Say". NBC News. 6 October 2022. Retrieved 26 September 2024.
  19. ^ "The dark side of ChatGPT". WJCL. 6 April 2023. Retrieved 26 September 2024.
  20. ^ "LOCAL IMPACT OF STEWARD HEALTH CLOSURES". WFMJ. 22 August 2024. Retrieved 26 September 2024.