Necdet Serhat Aybat
 |
Necdet Serhat Aybat |
Contact
360 Leonhard Bldg., University Park, PA 16802
Tel: (814) 867-1284
nsa10@psu.edu
https://nsaybat.org
Education
Columbia University, New York, Operations Research Ph.D. 2011
Columbia University, New York, Operations Research M.Phil. 2011
Bogazici University, Istanbul, Industrial Engineering M.S. 2005
Bogazici University, Istanbul, Industrial Engineering B.S. 2003
Research Interests
Methodology: convex optimization, first-order methods for large-scale problems, distributed optimization, constrained optimization techniques
Applications: compressed sensing, robust matrix decomposition, image and video processing, machine learning, power flow optimization on electricity grids
Editorial Board
Mathematics of Operations Research (May 2023 – )
Journal of Scientific Computing (December 2023 – )
Journal of Global Optimization (July 2024 – )
Grants
ONR Grant N00014-24-1-2666, “Collaborative Proposal: Primal-Dual Algorithms for Minimax Problems with Applications to Distributionally Robust Learning,” N. S. Aybat (PI, 50%) and Mert Gurbuzbalaban (PI, 50%), 09/01/24 - 08/31/27 ($800,000)
The Hal and Inge Marcus Funds (Penn State-Technion), “Efficient Methods for Nonconvex-Concave Minimax Problems,” N. S. Aybat (PI, 71%) and Shoham Sabach (PI, 29%), 08/01/24 - 07/31/26 ($34,000)
ONR Grant N00014-21-1-2271, “Collaborative Proposal: Robust Primal-Dual Algorithms for Saddle Point Problems with Applications to Multi-Agent Systems,” N. S. Aybat (PI, 50%) and Mert Gurbuzbalaban (PI, 50%), 04/01/21 - 03/31/24 ($652,000)
NSF Grant CMMI-1635106, “Decentralized power flow optimization on electricity grids via distributed consensus methods”, N. S. Aybat (PI), 09/01/2016 - 08/31/2020 ($235,852)
ARO Grant W911NF-17-1-0298, “Decentralized methods for multi-agent problems over networks,” N. S. Aybat (PI), 07/01/17 - 03/31/18 ($60,000)
NSF Grant CMMI-1400217, “Resolving Parametric Misspecification: Joint Schemes for Computation and Learning”, U. V. Shanbhag (PI, 50%), N. S. Aybat (co-PI, 50%), 08/01/2014 - 07/31/2017 ($300,000)
PhD Students
Qiushui Xu (PhD Student)
Xuan Zhang (PhD Student)
Erfan Yazdandoost Hamedani (PhD 2020), Tenure-Track Assistant Professor at University of Arizona
Ashkan M. Jasour (PhD 2016), Research Scientist at the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT
Zi Wang (PhD 2016), Senior Machine Learning Engineer at Instacart
Sam Davanloo Tajbakhsh (PhD 2015), Tenure-Track Assistant Professor at Ohio State
News
10/01/2024: A new preprint is out: “Adaptive Algorithms for Robust Phase Retrieval”
09/25/2024: The paper titled “High-probability Complexity Guarantees for Nonconvex Minimax Problems,” is accepted to NeurIPS 2024, Vancouver, Canada, December 11-13, 2024 (acceptance rate 25.8%).
06/20/2024: A new preprint is out: “AGDA+: Proximal Alternating Gradient Descent Ascent Method With a Nonmonotone Adaptive Step-Size Search For Nonconvex Minimax Problems”
05/30/2024: The paper titled “High Probability and Risk-Averse Guarantees for a Stochastic Accelerated Primal-Dual Method,” is accepted to be published in Journal of Machine Learning Research (JMLR).
05/23/2024: A new preprint is out: “High-probability Complexity Guarantees for Nonconvex Minimax Problems”
03/12/2024: A new preprint is out: “A Stochastic GDA Method With Backtracking For Solving Nonconvex (Strongly) Concave Minimax Problems”
01/16/2024: The paper titled “Lp Quasi-norm Minimization: Algorithm and Applications,” EURASIP Journal on Advances in Signal Processing.
12/08/2023: The paper titled “Jointly Improving the Sample and Communication Complexities in Decentralized Stochastic Minimax Optimization,” is accepted to the 38th Annual AAAI Conference on Artificial Intelligence.
09/03/2023: The paper titled “Robust Accelerated Primal-Dual Methods for Computing Saddle Points,” is accepted to be published in SIAM Journal on Optimization (SIOPT).
04/22/2023: The paper titled “A Fast Row-Stochastic Decentralized Optimization Method Over Directed Graphs,” is accepted to be published in IEEE Transactions on Automatic Control (TAC).
04/02/2023: The new preprint on high probability bounds for a primal-dual algorithm is available on arXiv: “High probability and risk-averse guarantees for stochastic saddle point problems”
02/10/2023: I will give a talk on “Stochastic Accelerated Primal-Dual Methods for Minimax Problems” at Industrial and Systems Engineering, Texas A&M
01/20/2023: The paper titled “Randomized Primal-Dual Methods with Line-Search for Saddle Point Problems,” is accepted to the 25th International Conference on Artificial Intelligence and Statistics (AISTATS), Valencia, Spain, April 25-27, 2023 (acceptance rate 29%).
01/09/2023: I gave a talk on “SAPD+: An Accelerated Stochastic Method for Nonconvex-Concave Minimax Problems” at US Mexico Workshop on Optimization and its Applications, Huatulco, Mexico, January 9-13, 2023.
09/14/2022: The paper titled “SAPD+: An Accelerated Stochastic Method for Nonconvex-Concave Minimax Problems,” is accepted to Advances in Neural Information Processing Systems (NIPS), New Orleans, USA, November 28 - December 9, 2022 (acceptance rate 25.6%, 2665 / 10411).
10/03/2021: The paper titled “A decentralized primal-dual method for constrained minimization of a strongly convex function” is accepted to be published in IEEE Transactions on Automatic Control (TAC).
09/12/2021: The paper titled “On the Analysis of Inexact Augmented Lagrangian Schemes for Misspecified Conic Convex Programs” is accepted to be published in IEEE Transactions on Automatic Control (TAC).
12/13/2020: The paper titled “A Primal-Dual Algorithm with Line Search for General Convex-Concave Saddle Point Problems” is accepted to be published in SIAM Journal on Optimization (SIOPT).
10/10/2020: The paper titled “On the Theoretical Guarantees for Parameter Estimation of Gaussian Random Field Models: A Sparse Precision Matrix Approach” is accepted to be published in Journal of Machine Learning Research (JMLR).
10/01/2020: I will give a talk on “A Primal-dual Algorithm With Linesearch For Nonbilinear Convex-concave Saddle Point Problems” at INFORMS Annual Metting 2020
07/01/2020: Aybat served in the 2020 INFORMS Computing Society Paper Prize Committee.
06/08/2018: The paper titled “Efficient Optimization Algorithms for Robust Principal Component Analysis and Its Variants” is accepted to be published in the Proceedings of the IEEE.
06/03/2018: I will give a talk on “An Accelerated Primal-dual Algorithm for General Convex-Concave Saddle Point Problems” at DIMACS Workshop on ADMM and Proximal Splitting Methods in Optimization
06/03/2018: I will give a talk on “An Accelerated Primal-dual Algorithm for General Convex-Concave Saddle Point Problems” at ISMP 2018
04/26/2018: The paper titled “End-to-End Distributed Flow Control for Networks with Nonconcave Utilities” is accepted to be published in IEEE Transactions on Network Science and Engineering.
02/15/2018: The symposium proposal tittled “Symposium on Distributed Learning and Optimization over Networks,” is accepted by IEEE GlobalSIP, Anaheim, California, USA, November 26-28. General chair: Zhi-Quan Luo; Technical co-chairs: N. S. Aybat, Mingyi Hong, Qing Ling. The Call for Papers (CFP) can be found here: https://2018.ieeeglobalsip.org/sym/18/DLN
01/09/2018: I gave a talk on “Multi-agent constrained optimization of a strongly convex function” at US Mexico Workshop on Optimization and its Applications, Huatulco, Mexico, January 8-12, 2018.
09/27/2017: The paper titled “Multi-agent constrained optimization of a strongly convex function over time-varying directed networks” is presented in the 55th Annual Allerton Conference on Communication, Control, and Computing (Allerton), Monticello, IL, 2017.
07/17/2017: The paper titled “Decentralized Computation of Effective Resistances and Acceleration of Consensus Algorithms,” is accepted to the 2017 5th IEEE Global Conference on Signal and Information Processing (GlobalSIP), Montreal, Canada, November 14-16, 2017.
07/17/2017: The paper titled “Multi-agent Constrained Optimization of a Strongly Convex Function,” is accepted to the 2017 5th IEEE Global Conference on Signal and Information Processing (GlobalSIP), Montreal, Canada, November 14-16, 2017.
07/03/2017: The paper titled “Generalized Sparse Precision Matrix Selection for Fitting Multivariate Gaussian Random Fields to Large Data Sets” is accepted to be published in Statistica Sinica.
06/28/2017: Aybat received Army Research Office (ARO) grant for the project titled “Decentralized methods for multi-agent problems over networks.”
05/21/2017: The paper titled “Non-Concave Network Utility Maximization: A Distributed Optimization Approach” received the IEEE INFOCOM 2017 “Best-in-Session-Presentation” award.
05/07/2017: The paper titled “Distributed Linearized Alternating Direction Method of Multipliers for Composite Convex Consensus Optimization” is accepted to be published in IEEE Transactions on Automatic Control.
05/05/2017: Aybat is promoted to the rank of associate professor with tenure in the College of Engineering at Penn State.
02/22/2017: I am orginizing the Nonlinear Optimization Cluster at the 2017 INFORMS annual meeting, Houston, TX, October 22-25, 2018. If you are interested in orginizing a session, please email me by March 15.
02/08/2017: The symposium proposal tittled “Symposium on Distributed Optimization and Resource Management over Networks,” is accepted by IEEE GlobalSIP, Montreal, Canada, November 14-16, 2017. General Co-chairs: Amir Asif, Zhi-Quan Luo; Technical co-chairs: N. S. Aybat, Mingyi Hong, Qing Ling.
01/21/2017: The paper titled “Distributed Non-Concave Network Utility Maximization in Connectionless Networks,” is accepted to the 2017 American Control Conference (ACC), Seattle, WA, USA, May 24-26, 2017.
11/25/2016: The paper titled “Non-Concave Network Utility Maximization: A Distributed Optimization Approach,” is accepted to IEEE International Conference on Computer Communications (INFOCOM), Atlanta, GA, USA, May 1-4, 2017 (acceptance rate 20.93%, 292 / 1395).
09/15/2016: Aybat is elected as the Vice Chair of Nonlinear Optimization SIG (Special Interest Group) of the INFORMS Optimization Society.
09/09/2016: Aybat's PhD student Zi Wang successfully defended his dissertation. He has joined Monsanto Company as an Operations Research Data Scientist in April 2016.
08/26/2016: Aybat's PhD student Ashkan Jasour (co-advised with Dr. Lagoa) successfully defended his dissertation, and will join the Dept. of Aeronautics and Astronautics at MIT as post-doctoral researcher in September 2016.
08/22/2016: Aybat's PhD student Sam Davanloo Tajbakhsh (co-advised with Dr. del Castillo) has joined the Dept. of Integrated Systems Engineering at the Ohio State University as a tenure-track assistant professor.
08/16/2016: Aybat is nominated for the Vice Chair for Nonlinear Optimization section of the INFORMS Optimization Society.
08/12/2016: The paper titled “A primal-dual method for constrained consensus optimization,” is accepted to Advances in Neural Information Processing Systems (NIPS), Barcelona, Spain, December 5-10, 2016 (acceptance rate 22.72%, 568 / 2500).
07/18/2016: The paper titled “Distributed primal-dual method for multi-agent sharing problem with conic constraints,” is accepted to the 50th Asilomar Conference on Signals, Systems and Computers, Pacific Grove, CA.