2024 ICS Student Paper Award Winner
The 2024 INFORMS Computing Society Student Paper Award winner is Shuvomoy Das Gupta (MIT) for the paper "Branch-and-Bound Performance Estimation Programming: A Unified Methodology for Constructing Optimal Optimization Methods," co-authored with Bart P.G. Van Parys, and Ernest K. Ryu.
The authors propose a methodology to design first-order methods for convex and nonconvex optimization problems, which provides an optimality guarantee in the sense that the resulting first-order method is provably the fastest first-order method for the problem at hand. Moreover, the proposed design mechanism is orders of magnitude faster than previous existing methods. The paper is clearly written, the technical content is thoroughly and elegantly explained, and the computational results are compelling and showcase the potential groundbreaking impact of the proposed approach. The committee was very impressed by this work and the decision to award it the first prize was unanimous.
The award committee also identified two runner-ups for the Award: Jie Wang (Georgia Tech) and Jiachang Liu (Duke University). You can find below the citation for each of their papers:
Runner-up: Jie Wang (Georgia Tech) for the paper "Variable Selection for Kernel Two-Sample Tests," co-authored with Santanu S. Dey, and Yao Xie
The paper considers a variable selection problem for two-sample tests and proposes mixed-integer programming formulations as well as approximation algorithms with performance guarantees. The authors masterfully present their contributions both from the optimization and statistical fields and report impressive computational results.
Runner-up: Jiachang Liu (Duke University) for the paper "OKRidge: Scalable Optimal k-Sparse Ridge Regression," co-authored by Chudi Zhong, Sam Rosen, and Cynthia Rudin.
The authors study a challenging quadratic optimization problem with indicators that has multiple applications in statistical learning. The authors propose a very creative fast lower bounding procedure based on a reformulation of the original problem as a saddle point optimization problem. The computational results show impressive performance for the propose method when compared to existing approaches.
The 2024 ICS Student Paper Award Committee members are:
- Beste Basciftci (University of Iowa)
- Austin Buchanan (Oklahoma State University)
- Leonardo Lozano, Chair (University of Cincinnati)
- Hamed Rahimian (Clemson University)
2023 ICS Student Paper Award Winner
The 2022 INFORMS Computing Society Student Paper Award winner is Shiyi Jiang (Hong Kong Polytechnic) for the paper "Optimized Dimensionality Reduction for Moment-based Distributionally Robust Optimization", arXiv:2305.03996, 2023, co-authored with Jianqiang Cheng (Arizona), Kai Pan (Hong Kong Polytechnic), Zuo-Jun Max Shen (Berkeley, Hong Kong).
In recognition of advancements in the field of distributionally robust optimization, the ICS Student Paper Award proudly presents this award to Shiyi Jiang and his co-authors for their pioneering research on the optimized dimensionality reduction (ODR) approach in moment-based distributionally robust optimization. Through an innovative integration of dimensionality reduction with subsequent optimization problems, the authors have illuminated new pathways for efficient problem-solving. Their development of modified ADMM algorithms has set a benchmark for tackling nonconvex SDPs. The demonstrable superiority of their method, reducing the computational time by magnitudes while maintaining solution precision, is a testament to their rigorous research and innovative application. The profound implications of their work on practical and theoretical fronts make them deserving recipients of this award.
Runner up: Mohammadhossein Mohammadisiahroudi (Lehigh) for the paper "An inexact feasible interior point method for linear optimization with high adaptability to quantum computers", arXiv:2307.14445, 2023, co-authored with Ramin Fakhimi, Zeguan Wu, Tamás Terlaky (Lehigh).
The 2023 ICS Student Paper Award Committee members are:
2022 ICS Student Paper Award Winner
The 2022 INFORMS Computing Society Student Paper Award winner is Irina Wang (Princeton) for the paper "Mean Robust Optimization," co-authored with Cole Becker, Bart Van Parys (MIT), and Bartolomeo Stellato (Princeton)
This paper introduces a framework that combines robust optimization with Wasserstein distributionally robust optimization. Notably, her algorithm provides an elegant mechanism to balance the trade-off between computational effort and conservatism. In a nutshell, the technique reduces the data points significantly (expediting computation) and carefully (preserving fundamental properties). The framework is accompanied by elegant theory giving finite-sample performance guarantees and conditions for which clustering does not increase conservatism and otherwise provides bounds on the effect on conservatism. The paper illustrates the efficiency and effectiveness of the framework with a thorough and meticulous computational study. The committee also praises the authors for making their code publically available.
Runner up: Peijing Liu (USC) for the paper "A graph-based decomposition method for convex quadratic optimization with indicators," Mathematical Programming, 2022, co-authored with Salar Fattahi (Michigan), Andrés Gómez (USC), Simge Küçükyavuz (Northwestern).
The 2022 ICS Student Paper Award Committee members are:
- Erick Moreno-Centeno, Chair (Texas A&M)
- Manish Bansal (Virginia Tech)
- Susan Hunter (Purdue)
- Alain Zemkoho (Southampton)
2021 ICS Student Paper Award Winner
The 2021 INFORMS Computing Society Student Paper Award winner is Aras Selvi for the paper "Convex Maximization via Adjustable Robust Optimization."
The paper presents a novel connection between (linear and nonlinear) convex maximization, which arises in a wide range of challenging theoretical and application domains, and adjustable robust optimization (ARO). This leads to a unified scheme that recasts convex maximization as an ARO model and produces lower and upper bounds using ARO techniques. This scheme applies to, for example, quadratic, geometric, and sum-of-max-linear-terms optimization models with multiple constraints. Extensive numerical results demonstrate that the proposed scheme can provide tight lower and upper bounds to multiple classes of convex maximization problems. In addition, it does so in computation time that is comparable to, and in many cases significantly shorter than, that consumed by state-of-the-art solvers.
Runner up: Yongchun Li for the paper "Beyond Symmetry: Best Submatrix Selection for the Sparse Truncated SVD."
The 2021 ICS Student Paper Award Committee members are:
- Merve Bodur (Toronto)
- Andres Gomez (USC)
- Ruiwei Jiang, Chair (Michigan)
- Gonzalo Munoz (Universidad de O’Higgins)
- Eunhye Song (Penn State)