Armin Eftekhari

I study the mathematical underpinnings of machine learning (a part of AI) and data science.

Contact

Works at

Assistant professor at Department of Mathematics and Mathematical Statistics

MIT-huset, plan 3, Matematik och matematisk statistik, MIT.D.327 Umeå universitet, 901 87 Umeå

I'm interested in problems within the intersection of statistics, optimization, and numerical analysis. For example, I often develop and analyze numerical algorithms that help machines extract information and learn from large amounts of data.

Read more about my research

Nonlinear matrix recovery using optimization on the Grassmann manifold

Applied and Computational Harmonic Analysis, Elsevier 2023, Vol. 62 : 498-542

Goyens, Florentin; Cartis, Coralia; Eftekhari, Armin

Over-parametrized matrix factorization in the presence of spurious stationary points

IEEE Transactions on Signal Processing, Institute of Electrical and Electronics Engineers (IEEE) 2022, Vol. 70 : 482-496

Eftekhari, Armin

Stable super-resolution of images: Theoretical study

Information and Inference, Oxford University Press 2021, Vol. 10, (1) : 161-193

Eftekhari, Armin; Bendory, Tamir; Tang, Gongguo

Explicit stabilised gradient descent for faster strongly convex optimisation

BIT Numerical Mathematics, Springer 2021, Vol. 61 : 119-139

Eftekhari, Armin; Vandereycken, Bart; Vilmart, Gilles; et al.

Subquadratic overparameterization for shallow neural networks

Advances in neural information processing systems (NeurIPS 2021), Curran Associates, Inc. 2021 : 11247-11259

Song, Chaehwan; Ramezani-Kebrya, Ali; Pethick, Thomas; et al.

Training linear neural networks: Non-local convergence and complexity results

37th International Conference on Machine Learning, ICML 2020, International Machine Learning Society (IMLS) 2020 : 2816-2827

Eftekhari, Armin

Double-loop unadjusted langevin algorithm

37th International Conference on Machine Learning, ICML 2020, International Machine Learning Society (IMLS) 2020 : 8139-8147

Rolland, Paul; Eftekhari, Armin; Kavis, Ali; et al.

Scalable Learning-Based Sampling Optimization for Compressive Dynamic MRI

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), IEEE 2020 : 8584-8588

Sanchez, Thomas; Gözcü, Baran; van Heeswijk, Ruud B.; et al.

View publications in DiVA

Research groups

Group member

Statistical Learning and Inference for Spatio-Temporal Data

Research projects

1 September 2019 until 31 August 2024

Compressive Sensing and Statistical Learning with Sparsity