Ryan Cory-Wright

Assistant Professor

Imperial Business School

About me

I am an Assistant Professor in the Analytics and Operations Group at Imperial Business School. My research focuses on optimization, machine learning, artificial intelligence, and their application to sustainability. I have received a Herman Goldstine Postdoctoral Fellowship (2022-23), the Nicholson Prize (2020), the Pierskalla Award (2020), the INFORMS DMDA Workshop Best Paper Award (2024), and have been a finalist in the M&SOM practice-based research competition (2023).

I am particularly interested in broadening the scope of optimization to address practical problems that current methods cannot solve to optimality. For instance, in my thesis, I have proposed a generalization of integer optimization that tackles rank constraints, which arise in product recommendation applications. Recently, I have worked on formulating the discovery phase of the scientific method as a convex optimization problem. I am also interested in leveraging optimization to support the transition to a low-carbon economy. For instance, we recently collaborated with OCP to develop a framework to guide a US$2 billion investment in solar panels and batteries.

Before joining IB, I spent a year as a postdoctoral fellow at IBM Research (Cambridge, MA). I received my PhD in Operations Research from MIT in 2022, advised by Dimitris Bertsimas. Before joining MIT, I received a BE (first class honors) in Engineering Science from the University of Auckland.

Prospective PhD students: I am always recruiting motivated and talented PhD students with a strong background in mathematics, operations, computer science, economics, statistics, or similar to join my research group. PhD admissions are handled at the department level. To work with me, please apply to Imperial Business School's PhD program in Analytics and Operations, specifying my name in the application and the types of research problems you find exciting.

News

October 2025: Overhauled Improved Approximation Algorithms for Low-Rank Problems. We have an approximation guarantee for orthogonal quadratic optimization, logarithmic in the number of columns.

September 2025: New preprint out on AI Noether - Bridging the Gap Between Scientific Laws Derived by AI Systems and Canonical Knowledge via Abductive Inference.

September 2025: New preprint out on The Need for Verification in AI-Driven Scientific Discovery.

August 2025: Excited to be serving as a track chair for IOS 2026 in Atlanta!

August 2025: Excited to be serving on Imperial Business School's cross-departmental AI Committee.

May 2025: New preprint out on Stability Regularized Cross-Validation.

May 2025: Excited to serve on the 2025 ICS student paper award committee with Austin Buchanan (chair), Yongchun Li, and Young-Woong Park. PhD students: please submit your best eligible work. Full details here.

April 2025: It was great to co-organize the 2025 London Operations Research Day (LORD 2025) with Jean Pauphilet , Agni Orfanoudaki and Yi-Chun Akchen. Check out a live thread of the conference on Bluesky.

October 2024: Our paper A Stochastic Benders Decomposition Scheme for Large-Scale Stochastic Network Design has been accepted for publication by INFORMS Journal on Computing.

June 2024: Our paper AI Hilbert has been accepted by Nature Communications and recognized with an Outstanding Technical Achievement Award by IBM. Check out a behind-the-paper blogpost.

April 2024: It was great to co-organize the 2024 London Operations Research Day (LORD 2024) with Jean Pauphilet and Agni Orfanoudaki . Check out a live thread of the conference on Twitter.

March 2024: Organizing session FriB8 at IOS 2024. I will be talking about our paper AI Hilbert and Andrés Gómez will be talking about our paper A New Approach to Cross-Validation for Sparse Regression.

January 2024: Delighted to be selected as a 2024 INFORMS Journal on Computing Meritorious Reviewer.

January 2024: Excited to be teaching classes on Decision Making Under Uncertainty (PhD), Optimization and Decision Models (Online MSc Business Analytics), and Data Structures and Algorithms (UG Economics, Finance, and Data Science). The PhD class is new—check out the syllabus.

November 2023: Our paper on Decarbonizing OCP has been accepted for publication at M&SOM as part of the 2023 M&SOM practice-based research competition.

November 2023: Excited to be giving talks on our paper “Disjunctive Branch-and-Bound for Certifiably Optimal Low-Rank Matrix Completion” at the Toronto Rotman Young Scholar Seminar Series (virtual) and the Imperial Control and Optimization Seminar Series (in-person, slides).

November 2023: Overhauled our paper Sparse PCA With Multiple Components, including adding a new section and open-sourcing the code. Please do check out the new version!

October 2023: Our paper Sparse Plus Low Rank Matrix Decomposition: A Discrete Optimization Approach has been accepted for publication at The Journal of Machine Learning Research.

October 2023: I started co-advising Lingjun Meng (joint with Wolfram Wiesemann) as a doctoral student.

August 2023: New preprint out on Evolving Scientific Discovery by Unifying Data and Background Knowledge with AI Hilbert.

July 2023: Officially joined the Analytics and Operations group at Imperial Business School.

June 2023: New preprint out on Stability-Adjusted Cross-Validation for Sparse Linear Regression . See thread on Twitter!

June 2023: Our paper on Decarbonizing OCP has been selected as a finalist in the 2023 M&SOM practice-based research competition, and as an honorable mention in the 2023 MIT Operations Research Center Student Paper Competition. I will be presenting this at M&SOM 2023 (slides) .

May 2023: New preprint out on Disjunctive Branch-and-Bound for Certifiably Optimal Low-Rank Matrix Completion. I will be presenting this at the MIP Workshop.

December 2022: Our paper A New Perspective on Low-Rank Optimization has been accepted for publication at Mathematical Programming.

March 2022: Defended my thesis Integer and Matrix Optimization: A Nonlinear Approach. Check out the five-page summary.

March 2022: Honored to be selected as a Herman Goldstine Postdoctoral Fellow!

February 2022: I am joining Imperial Business School starting in July 2023! Very much looking forward to new collaborations and projects at Imperial.

January 2022: Our paper Solving Large-Scale Sparse PCA to Certifiable (Near) Optimality has been accepted for publication at the Journal of Machine Learning Research.

October 2021: Excited to receive the 2021 INFORMS Data Mining Student Paper Award with Nicholas Johnson.

August 2021: Our paper A Scalable Algorithm for Sparse Portfolio Selection has been accepted for publication at INFORMS Journal on Computing.

July 2021: Our paper Mixed-Projection Conic Optimization: A New Paradigm for Modeling Rank Constraints has been accepted for publication at Operations Research.

November 2020: Very excited to receive the 2020 INFORMS George Nicholson Prize with Jean Pauphilet.

Working Papers

20. AI Noether: Bridging the Gap Between Scientific Laws Derived by AI Systems and Canonical Knowledge via Abductive Inference
K. Srivastava, S. Dash, R. Cory-Wright, B. Trager, and L. Horesh, Submitted
Preprint Code

19. Stability Regularized Cross-Validation
R. Cory-Wright and A. Gómez, Submitted
Preprint

18. Improved Approximation Algorithms for Low-Rank Problems Using Semidefinite Optimization
R. Cory-Wright and J. Pauphilet, Submitted
Preprint

17. Disjunctive Branch-and-Bound for Certifiably Optimal Low-Rank Matrix Completion
D. Bertsimas, R. Cory-Wright, S. Lo, and J. Pauphilet, Submitted
Preprint Slides Code

16. Optimal Cross-Validation for Sparse Linear Regression
R. Cory-Wright and A. Gómez, Submitted
Major Revision, INFORMS Journal on Computing
Preprint

15. Sparse PCA With Multiple Components
R. Cory-Wright and J. Pauphilet
Major Revision, Operations Research
First place, 2024 INFORMS DMDA Workshop Paper Award (theoretical track)
Preprint Slides Code

14. The Need for Verification in AI-Driven Scientific Discovery
C. Cornelio, T. Ito, R. Cory-Wright, S. Dash, and L. Horesh, Submitted
Minor Revision, Philosophical Transactions of the Royal Society A
Preprint

Journal Papers

13. Decarbonizing OCP
D. Bertsimas, R. Cory-Wright, and V. Digalakis Jr.
Manufacturing & Service Operations Management, 2025+
Finalist, 2023 Manufacturing & Service Operations Management Practice-Based Research Competition
Honorable Mention, 2023 MIT ORC Student Paper Competition (Digalakis)
Preprint DOI Slides OCP's green investment program Imperial business article

12. A Stochastic Benders Decomposition Scheme for Large-Scale Stochastic Network Design
D. Bertsimas, R. Cory-Wright, J. Pauphilet, and P. Petridis
INFORMS Journal on Computing, 37(5):1163-1181, 2025
Preprint DOI

11. Evolving Scientific Discovery by Unifying Data and Background Knowledge with AI Hilbert
R. Cory-Wright, C. Cornelio, S. Dash, B. El Khadir, and L. Horesh
Nature Communications, 15, 5922, 2024
Meet AI Hilbert, a new algorithm for transforming scientific discovery (IBM Research Blogpost)
Preprint DOI Code Website Behind the paper Slides Seminar recording

10. Sparse Plus Low-Rank Matrix Decomposition: A Discrete Optimization Approach
D. Bertsimas, R. Cory-Wright, and N. A. G. Johnson
Journal of Machine Learning Research, 24(267):1-51, 2023
First place, 2021 INFORMS Data Mining Student Paper Award
Preprint DOI Code

9. A New Perspective on Low-Rank Optimization
D. Bertsimas, R. Cory-Wright, and J. Pauphilet
Mathematical Programming, 202:47-92, 2023
Preprint DOI Slides Code Video

8. Mixed-Projection Conic Optimization: A New Paradigm for Modeling Rank Constraints
D. Bertsimas, R. Cory-Wright, and J. Pauphilet
Operations Research, 70(6):3321-3344, 2022
First place, 2020 INFORMS George Nicholson Paper Award
Preprint DOI Code Video Poster

7. Solving Large-Scale Sparse PCA to Certifiable (Near) Optimality
D. Bertsimas, R. Cory-Wright, and J. Pauphilet
Journal of Machine Learning Research, 23(13):1-35, 2022
Preprint DOI Code Slides

6. A Scalable Algorithm for Sparse Portfolio Selection
D. Bertsimas and R. Cory-Wright
INFORMS Journal on Computing, 34(3): 1489-1511, 2022
Preprint DOI Code Poster

5. A Unified Approach to Mixed-Integer Optimization Problems With Logical Constraints
D. Bertsimas, R. Cory-Wright, and J. Pauphilet
SIAM Journal on Optimization, 31(3):2340-2367, 2021
First place, 2019 INFORMS Computing Society Student Paper Award
Preprint DOI ICS newsletter Video

4. From Predictions to Prescriptions: A Data-Driven Response to COVID-19
D. Bertsimas et al.
Health Care Management Science, 24:253-272, 2021
First place, 2020 INFORMS Health Applications Society Pierskalla Paper Award
Preprint DOI Code NYT

3. On Stochastic Auctions in Risk-Averse Electricity Markets With Uncertain Supply
R. Cory-Wright and G. Zakeri
Operations Research Letters, 48(3):376-384, 2020
Preprint DOI

2. On Polyhedral and Second-Order Cone Decompositions of Semidefinite Optimization Problems
D. Bertsimas and R. Cory-Wright
Operations Research Letters, 48(1):78-85, 2020
Preprint DOI

1. Payment Mechanisms for Electricity Markets With Uncertain Supply
R. Cory-Wright, A. Philpott, and G. Zakeri
Operations Research Letters, 46(1):116-121, 2018
First place, 2016 ORSNZ Young Practitioner’s Prize
Preprint DOI

Integer and Matrix Optimization: A Nonlinear Approach
R. Cory-Wright. Advised by: D. Bertsimas
Ph.D. Thesis, Massachusetts Institute of Technology, May 2022
MIT libraries Five-page summary

Teaching

At Imperial:

ELEC70137 Fundamentals of Python (MSc AI Applications and Innovation, Fall 2025)
A course that introduces the fundamentals of Python

ELEC70137 Python and Machine Learning (MSc AI Applications and Innovation, Fall 2024)
Designed a new course introducing machine learning in Python

BUSI40008 Data Structures and Algorithms (UG Economics, Finance and Data Science, Spring 2024, Spring 2025, Spring 2026)
Designed a new course introducing computational problem-solving through the lens of algorithms and data structures in Python

BUSI70447 Decision Making Under Uncertainty (PhD, Spring 2024, Spring 2025, Spring 2026)
Designed a new course introducing techniques for decision-making under uncertainty widely used in operations research. Includes stochastic optimization, robust optimization, and dynamic programming Lecture 1 Slides Lecture 2 Slides Lecture 3 Slides

BUSI70218 Optimization and Decision Models (Online MSc Business Analytics, Spring 2024)
A course that introduces students to theory and applications of linear, discrete, and nonlinear optimization

At MIT:

Head TA, 15.095 Machine Learning Under a Modern Optimization Lens (MBAn/PhD, Fall 2019, 2021)
Head TA, 15.071 The Analytics Edge (MBA, Fall 2020)
TA, 15.093 Optimization Methods (MSc/MBAn/PhD, Fall 2018)
Participant, Kaufman Teaching Certificate Program (Fall 2021)
Session Instructor, 15.S60 Computing for Optimization and Statistics (MSc/PhD, Jan 2019, 2020)