Publications
Papers in nonsmooth optimization, variational analysis, and inverse problems, with applications to PDE-constrained optimization and machine learning. Numbers follow my CV.
Journal articles
-
[1]
Nonsmooth Newton methods with effective subspaces for polyhedral regularization
Mathematics of Operations Research, to appear (accepted September 2026)
Newton-type methods for problems with polyhedral regularizers (Lasso, ℓ∞, OSCAR/SLOPE, total variation) that use only first-order information about the regularizer and converge quadratically under tilt stability.
-
[2]
Stable recovery of regularized linear inverse problems
Inverse Problems 41(6), 065018 (2025)
A complete characterization, using second-order variational analysis, of when a structured signal can be recovered from noisy measurements with error proportional to the noise.
-
[3]
A regularity result via fractional maximal operators for p-Laplace equations in weighted Lorentz spaces
Complex Variables and Elliptic Equations 67(7), 1737–1755 (2022)
-
[4]
A gradient estimate related fractional maximal operators for a p-Laplace problem in Morrey spaces
Taiwanese Journal of Mathematics 25(4), 809–829 (2021)
Submitted
-
[5]
A globally convergent Newton method for nonsmooth optimization in the absence of smooth surrogates
Submitted to Journal of Scientific Computing (2026)
-
[6]
Isolated calmness of regularized linear inverse problems
Submitted to Optimization (2026)
When does the solution of a regularized inverse problem move at most linearly with the data? A tangent-cone characterization, with checkable tests based on uniqueness and linear programming.
-
[7]
Inexact DC algorithms in Hilbert spaces with applications to PDE-constrained optimization
Submitted to Applied Mathematics & Optimization (2026)
An inexact, accelerated DC algorithm in Hilbert spaces with convergence guarantees, applied to sparse elliptic optimal control with a nonconvex L1−L2 penalty.
-
[8]
Consistently informative soft-label temperature for knowledge distillation
Submitted to the International Conference on Artificial Intelligence and Statistics (AISTATS 2027)
Per-sample temperatures for teacher and student in knowledge distillation, so every soft label carries a similar amount of information.
Preprints
- [9]
- [10]
- [11]