Welcome to my website. My name is Albert S. Berahas and I am an Assistant Professor in the Department of Industrial and Operations Engineering at the University of Michigan.
My research focuses on designing, developing, analyzing and implementing algorithms for solving large scale nonlinear optimization problems. Such problems are ubiquitous, and arise in a plethora of areas such as engineering design, economics, transportation, robotics, machine learning and statistics. Specifically, I am interested in and have explored several sub-fields of nonlinear optimization such as: (i) general nonlinear optimization algorithms, (ii) optimization algorithms for machine learning, (iii) constrained optimization, (iv) stochastic optimization, (v) derivative-free optimization, and (vi) distributed optimization. I am affiliated with the Michigan Institute for Data Science (MIDAS) and with the Michigan Institute for Computational Discovery and Engineering (MICDE).
Email: aberahas@umich.edu, albertberahas@gmail.com
Office: 2783 IOE Building, 1205 Beal Avenue Ann Arbor, MI 48109 (map)
Quick Links: Full CV, Bio, Google Scholar, arxiv, LinkedIn
Announcements
I am looking for outstanding and highly motivated PhD students and Postdoctoral Research Fellows to join my team. If interested, please send me an email with your CV, and apply to our PhD program.
News
- September 2021: Invited to be the chair of the Nonlinear Optimization cluster for the INFORMS 2022 Optimization Society Conference
- August 2021: Our paper “Quasi-Newton Methods for Machine Learning: Forget the Past, Just Sample“ (joint work with Majid Jahani, Peter Richtárik and Martin Takáč) has been accepted for publication in Optimization Methods & Software
- August 2021: New Paper “Full-low evaluation methods for derivative-free optimization“ (joint work with Frank E. Curtis, Michael J. O’Neill and Daniel P. Robinson)
- August 2021: Chair of the organizing committee of the MOPTA 2021 conference August 2-4, 2021
- July 2021: Co-organized the “Beyond First Order Methods in Machine Learning Systems“ workshop at ICML 2021 (co-organizers Raghu Bollapragrada, Rixon Crane, Amir Gholami, J. Lyle Kim, Anastasios (Tasos) Kyrillidis, and Michael Mahoney, Fred Roosta, Rachael Tappenden)
- July 2021: New Paper “A Stochastic Sequential Quadratic Optimization Algorithm for Nonlinear Equality Constrained Optimization with Rank-Deficient Jacobians“ (joint work with Frank E. Curtis, Michael J. O’Neill and Daniel P. Robinson)
- July 2021: Presented in a session organized by Thomas O’Leary-Roseberry, Peng Chen and Omar Ghattas (Beyond First Order Methods in Large-Scale Stochastic Optimization) at the SIAM Optimization Conference
- July 2021: Organized six sessions (Methods for Large-Scale, Nonlinear and Stochastic Optimization I/II/III and Recent Advancements in Optimization Methods for Machine Learning I/II/III) at the SIAM Optimization Conference
- July 2021: Presented in a session organized by Clément Royer (Derivative-free Optimization and Connections to Machine Learning) at the 31st European Conference on Operational Research (EURO 2021)
- July 2021: Organized three sessions (Beyond first order methods in machine learning) and presented at the 18th International Workshop on Continuous Optimization (EUROPT 2021)
- June 2021: Our paper “On the Convergence of Nested Decentralized Gradient Methods with Multiple Consensus and Gradient Steps“ (joint work with Raghu Bollapragada and Ermin Wei) has been accepted for publication in IEEE Transactions on Signal Processing
- May 2021: Our paper “Sequential Quadratic Optimization for Nonlinear Equality Constrained Stochastic Optimization“ (joint work with Frank E. Curtis, Daniel P. Robinson and Baoyu Zhou) has been accepted for publication in SIAM Journal on Optimization
- May 2021: Our paper “Auction-Based Preferential Shift Scheduling: A Case Study on the Lehigh University Libraries“ (joint work with Sudeep Metha and Ved Patel) has been accepted for publication in the Institute of Industrial and Systems Engineers Annual Conference and Expo, IISE 2021
- April 2021: Presented in the Statistics Seminar at UW Madison (Department of Statistics)
- April 2021: Presented in a session organized by Giampaolo Luizzi (Optimization in Machine Learning) at the 3rd IMA and OR Society Conference on Mathematics of Operational Research (IMA ORS)
- April 2021: Our ICML 2020 workshop “Beyond First Order Methods in Machine Learning Systems“ (co-organizers Raghu Bollapragrada, Rixon Crane, Amir Gholami, J. Lyle Kim, Anastasios (Tasos) Kyrillidis, and Michael Mahoney, Fred Roosta, Rachael Tappenden) has been accepted (all past workshops can be found here)
- March 2021: Our paper “A Theoretical and Empirical Comparison of Gradient Approximations in Derivative-Free Optimization“ (joint work with Liyuan Cao, Krzysztof Choromanski and Katya Scheinberg) has been accepted for publication in Foundations of Computational Mathematics
- March 2021: Our paper “Global Convergence Rate Analysis of a Generic Line Search Algorithm with Noise“ (joint work with Liyuan Cao and Katya Scheinberg) has been accepted for publication in SIAM Journal on Optimization
- March 2021: Presented in a session organized by Amir Gholaminejad, Fred Roosta and Michael Mahoney (Beyond First Order Methods in Machine Learning Systems) at SIAM CSE 2021
- February 2021: Organized the IOE department seminars this semester
- February 2021: Our paper “SONIA: A Symmetric Blockwise Truncated Optimization Algorithm“ (joint work with Majid Jahani, Mohammadreza Nazari, Rachael Tappenden and Martin Takáč) has been accepted for publication at the 24th International Conference on Artificial Intelligence and Statistics (AISTATS 2021)
- January 2021: Our paper “Limited-Memory BFGS with Displacement Aggregation“ (joint work with Frank E. Curtis and Baoyu Zhou) has been accepted for publication in Mathematical Programming
(For the full list of news please check the News tab.)