About Me

I am Valentio Iverson, a first-year PhD student in Electrical Engineering and Computer Science (EECS) at the Massachusetts Institute of Technology (MIT), where I work on the theoretical foundations of machine learning. I am fortunate to be advised by Gabriele Farina, and I am currently working with him on understanding lower bounds in online learning problems. I completed my Bachelor of Mathematics at the University of Waterloo in Computer Science, Pure Mathematics, and Combinatorics & Optimization, with a minor in Computational Mathematics.

My main research interests are in machine learning theory, especially questions about learning in adaptive environments and deep learning theory. Broadly, I am interested in developing provable algorithms and lower bounds for simple structures or phenomena that help explain how modern learning systems behave.

I am also active in the mathematical problem-solving community. I wrote Problem 2 at IMO 2024, the first IMO problem from Indonesia, and received an Honorable Mention in the 2024 William Lowell Putnam Mathematical Competition.

For contest handouts, problem proposals, coursework notes, and related materials, see my Math page. For expository writing on mathematics and machine learning, visit my Blog Posts page.

My Erdos Number is 3.

Research

My research is broadly in machine learning theory, with current interests in online learning and learning in adaptive environments and deep learning theory. I have also previously worked on sampling and statistical estimation and classical learning theory. Across these areas, I am especially interested in provable algorithms, lower bounds, and simple structural explanations for how learning systems behave.

For more details, see my Research & Publications page.

Selected Publications & Manuscripts

• Robust Statistical Estimators with Bounded Empirical Sensitivity
  Valentio Iverson, Gautam Kamath, Argyris Mouzakis, Adam Smith.
  Neural Information Processing Systems (NeurIPS) 2026

• On the Importance of Randomization in Discriminative Feature Feedback
  Valentio Iverson, Tosca Lechner, Sivan Sabato.
  Conference on Learning Theory (COLT) 2026

• Optimal Differentially Private Sampling of Unbounded Gaussians
  Valentio Iverson, Gautam Kamath, Argyris Mouzakis.
  Conference on Learning Theory (COLT) 2025 and Theory and Practice of Differential Privacy (TPDP) 2025.

• Mean and Variance Estimation Complexity in Arbitrary Distributions via Wasserstein Minimization
  Valentio Iverson, Stephen Vavasis.
  Preprint; submitted to Journal of Global Optimization.

News

• September 24, 2026: Robust Statistical Estimators with Bounded Empirical Sensitivity was accepted to NeurIPS 2026.
• September 9, 2026: Officially started my PhD in MIT EECS. Pretty excited for what’s to come!
• September 3, 2026: On a Conjecture about the 3-adic Valuation of a Binomial Sum was accepted to the International Journal of Number Theory.
• June 11, 2026: Multiplicatively Dependent Integer Vectors on a Hyperplane was accepted to the Journal of the London Mathematical Society.

Selected Honors & Awards

• Alan V. Oppenheim Fellowship, Massachusetts Institute of Technology, 2026.
• CRA Outstanding Undergraduate Researcher Award (Honorable Mention), Computing Research Association, 2026.
• Nathan Mendelsohn Prize, University of Manitoba, 2025.
• Germain-Erdős Undergraduate Research Award, University of Waterloo, 2025.
• Honorable Mention and member of the University of Waterloo Putnam Team, William Lowell Putnam Mathematical Competition, 2024.
• Problem Author (Problem 2), International Mathematical Olympiad, 2024.
• Bronze Medal, Asia Pacific Mathematics Olympiad, 2020.
• Absolute Winner, Indonesia National Mathematical Olympiad, 2019.

Contact

Feel free to reach out for collaborations or inquiries: 📧 viverson@mit.edu