Damek Davis

Associate Professor of Statistics and Data Science

Contact Information

Primary Email:
damek@wharton.upenn.edu

office Address:
319 Academic Research Building
265 South 37th Street
Philadelphia, PA 19104

Research

Xingyuang Lu, Minh Pham, Elisa Negrini, Damek Davis, Stanley Osher, Jianwei Miao (2024), Computational Microscopy beyond Perfect Lenses, Physical Review E, 110 (5).
Damek Davis, Dmitriy Drusvyatskiy, Liwei Jiang (2024), Asymptotic normality and optimality in nonsmooth stochastic approximation, The Annals of Statistics , 52 (4), pp. 1485-1508.
Jeongyeol Kwon, Wei Qian, Constantine Caramanis, Yudong Chen, Damek Davis, Nhat Ho (2024), Global Optimality of the EM Algorithm for Mixtures of Two Linear Regression, IEEE Transactions on Information Theory, 70 (9), pp. 6519-6546.
Damek Davis and Liwei Jiang (2024), A local nearly linearly convergent first-order method for nonsmooth functions with quadratic growth, Foundations of Computational Mathematics .
Damek Davis, Dmitriy Drusvyatskiy, Vasileios Charisopoulos (2023), Stochastic algorithms with geometric step decay converge linearly on sharp functions, Mathematical Programming, 207 (p.p. 145-190).
Vasileios Charisopoulos and Damek Davis (2023), A superlinearly convergent subgradient method for sharp semismooth problems, Mathematics of Operations Research , 49 (3), pp. 1678-1709.
Damek Davis (2023), Variance reduction for root-finding problems, Mathematical Programming, 197 (p.p. 375-410).
Damek Davis, Mateo Diaz, Dmitriy Drusvyatskiy (2022), Escaping strict saddle points of the Moreau envelope in nonsoomth optimization, SIAM Journal on Optimization , 32 (3), pp. 1958-1983.
Damek Davis and Dmitriy Drusvyatskiy (2022), Conservative and semismooth derivatives are equivalent for semialgebraic maps, Set-Valued and Variational Analysis , 30 (p.p. 453 -463).
Damek Davis, Mateo Diaz, Kaizheng Wang (Forthcoming), Clustering a Mixture of Gaussians with Unknown Covariance.
All Research from Damek Davis »

Teaching

Current Courses

STAT4830 - Numerical Optimization For Data Science And Machine Learning
Optimization is the modeling language in which modern data science, machine learning, and sequential decision-making problems are formulated and solved numerically. This course will teach students how to formulate these problems mathematically, choose appropriate algorithms to solve them, and implement and tune the algorithms in the software PyTorch. PyTorch is a freely available machine learning library is recognized as one of the two most popular machine learning libraries alongside TensorFlow. By the end of the course, students will become an intelligent consumer of numerical methods and software for solving modern optimization problems.
STAT4830001 ( Syllabus )
STAT4830002 ( Syllabus )

Past Courses

STAT4830 - Optimization for Data Sci & ML
Optimization is the modeling language in which modern data science, machine learning, and sequential decision-making problems are formulated and solved numerically. This course will teach students how to formulate these problems mathematically, choose appropriate algorithms to solve them, and implement and tune the algorithms in the software PyTorch. PyTorch is a freely available machine learning library is recognized as one of the two most popular machine learning libraries alongside TensorFlow. By the end of the course, students will become an intelligent consumer of numerical methods and software for solving modern optimization problems.

Activity

Latest Research

Xingyuang Lu, Minh Pham, Elisa Negrini, Damek Davis, Stanley Osher, Jianwei Miao (2024), Computational Microscopy beyond Perfect Lenses, Physical Review E, 110 (5).

All Research