QUOC TRAN-DINH + Department of Statistics and Operations Research + Adjunct Member of Department of Computer Science (2020-present) Address: 333 Hanes Hall CB #3260 The University of North Carolina at Chapel Hill Chapel Hill, NC 27599 E-mail: quoctdATemailDOTuncDOTedu Phone: +1-919-843-6023 Name in native language: Trần Đình Quốc (Quốc is my first name).
- October 30: Our paper A new homotopy proximal variable-metric framework for composite convex minimization gets accepted for publication on Mathematics of Operations Research (This is joint work with Ling Liang and Kim-Chuan Toh (NUS, Singapore)).
- October 12: Our paper A Hybrid Stochastic Optimization Framework for Stochastic Composite Nonconvex Optimization gets accepted for publication on Mathematical Programming, Ser. A.
- September 29: Our paper Hybrid Variance-Reduced SGD Algorithms for Nonconvex-Concave Minimax Problems (Q. Tran-Dinh, Deyi Liu, and Lam M. Nguyen) got accepted for presenting at NeurIPs 2020.
- September 29: I serve as an Associate Editor of Computational Optimization and Applications (COAP).
- September 10: I have been appointed for an adjunct position at the CS department.
- August 1, 2020: Our paper: Non-Stationary First-Order Primal-Dual Algorithms with Fast Convergence Rates has been accepted for publication on SIAM J. Optimization.
- June 1, 2020: Our paper: Stochastic Gauss-Newton Algorithms for Nonconvex Compositional Optimization (Joint work with N. H. Pham, and L. M. Nguyen) has been accepted for presenting at ICML 2020.
- May 5, 2020: Our paper: ProxSARAH: An efficient algorithmic framework for stochastic composite nonconvex optimization has been accepted for publication on Journal of Machine Learning Research.
- February 3, 2020: Our paper (with my student and colleague) has been accepted for publication on COAP (Computational Optimization and Applications).
- January 28, 2020: A standard ONR grant gets awarded for 3 years.
- January 7, 2020: Our paper: A Hybrid Stochastic Policy Gradient Algorithm for Reinforcement Learning (N. H. Pham, L. M. Nguyen, D. T. Phan, H. P. Nguyen, M. van Dijk, and Q. Tran-Dinh) got accepted for publication in the proceedings of the 23rd International Conference on Artificial Intelligence and Statistics (AISTATS 2020), Italy, 2020.
- January 6, 2020: I am teaching STOR 614: Linear Programming and Extensions this Spring. The course covers theory, solution methods, and applications of LPs, QPs, second-order cone programming, semidefinite programming, and convex programming. If you are interested in, please take this course. The tentative syllabus can be found HERE.
- December 23, 2019: Our paper Transferring Optimality Across Data Distributions via Homotopy Methods (Matilde Gargiani, Andrea Zanelli, Quoc Tran-Dinh, Moritz Diehl, Frank Hutter) got accepted for publication in the proceedings of the Eighth International Conference on Learning Representations (ICLR 2020), Ethiopia, April 2020.
- August 3-8, 2019: I am attending ICCOPT 2019 in Berlin. Together with I. Necoara, we organize a section: “Recent advances in first-order methods for constrained convex optimization and related problems” with 4 parts running from Monday to Tuesday with 12 talks (9 invited talks and 3 contributed talks). Please join us!
- I will be taking a research leave in Fall 2019, and plan to visit some universities in the US and Europe. I am happy to connect to researchers who have common research interests as I do. I will also be a SAMSI fellow within the Deep Learning Program in Fall 2019.
- March 15, 2019: We have uploaded a new preprint: “Non-Stationary First-Order Primal-Dual Algorithms with Fast NonErgodic Convergence Rates” on Arxiv (https://arxiv.org/pdf/1903.05282.pdf). This paper is concerned with faster than O(1/k) convergence rates of primal-dual methods. Any feedback is highly appreciated.
- February 15, 2019: Together with my student, we have completed a manuscript: ProxSARAH: An Efficient Algorithmic Framework for Stochastic Composite Nonconvex Optimization. This is a joint work with Lam Nguyen M. and Dzung Phan T. at IBM Waston. The preprint can be found at: https://arxiv.org/pdf/1902.05679.pdf. Any comments and suggestions are highly appreciated.
- August 21st, 2018: My first student (co-advisor with Shu), Tianxiao Sun, has successfully finished his Ph.D. defense on August 20, 2018. He is going to work at Lowe’s in September. I am still looking for a new well-motivated Ph.D. student to work with me in the area of numerical optimization: theory and algorithms (with possible applications in machine learning and data analysis).
- April 30, 2018: Our paper “Generalized self-concordant functions: A recipe for Newton-type methods” has been accepted for publication on Mathematical Programming. The preprint version is on arxiv.org (https://arxiv.org/pdf/1703.04599.pdf [PDF]).
- April 1, 2018: In Fall 2018, I am teaching a special topic course: Selected Numerical Methods for Modern Optimization in Data Analysis. The tentative syllabus can be found HERE. Please join us!!!! Some lecture notes are posted online HERE.
- March 30, 2018: Our paper “Self-concordant inclusions: A unified framework for path-following generalized Newton-type algorithms” has been published in Mathematical Programming. The preprint can be found at https://arxiv.org/pdf/1707.07403.pdf.
- November 30, 2017: I have completed a manuscript: “Proximal Alternating Penalty Algorithms for Nonsmooth Constrained Convex Optimization“. This paper aims at developing first-order methods for solving general constrained convex optimization problems. The MATLAB code is available at https://github.com/quoctd/PAPA-1.0. If you find any bug, please let me know.
- September 19, 2017: After almost two years, our paper “A single-phase, proximal path-following framework“ has been accepted for publication on Mathematics of Operations Research.
- September 4, 2017: Our conference paper “Smooth Primal-Dual Coordinate Descent Algorithms for Nonsmooth Convex Optimization” has been accepted at NIPS, 2017, in Long Beach, CA. I am planning to attend it in December.
- September 1, 2017: Ahmet Alacaoglu, a Ph.D. student from LIONS, EPFL, Switzerland, is visiting me as an internship student. He will be here until September 30, 2017, and is working on our joint project.
- August 10, 2017: I have just completed a paper: “A new inexact homotopy proximal Newton algorithms for scalar parametric composite convex minimization. It deals with an interesting topic: Homotopy second-order methods. I have not posted this paper on Arxiv, but if you are interested in, please let me know.
- August 7, 2017: Prof. Ion Necoara, a Fulbright scholar from the University Politehnica Bucharest, Romania, is currently visiting me at UNC-Chapel Hill. He is with our department until the end of November 2017.
- July 27, 2017: After a long time, our paper “A smooth primal-dual optimization framework for nonsmooth composite convex minimization” has been accepted for publication on SIAM J. Optimization. The preprint can be found at https://arxiv.org/abs/1507.06243 [pdf].
- July 26, 2017: I gave a talk on “Exploiting analytical structures in convex optimization applications” at the IBM Waston Research Center.
- July 15, 2017: With my student, Tianxiao Sun, and my colleague, Shu Lu, we have posted our paper “Self-concordant inclusions: A unified framework for path-following generalized Newton-type algorithms” on Arvix (https://arxiv.org/pdf/1707.07403.pdf). This paper was completed for a year ago and is under revision [pdf].
- March 14, 2017: Together with my student, Tianxiao Sun, we have posted a new paper “Generalized self-concordant functions: A recipe for Newton-type methods” on arxiv.org (https://arxiv.org/pdf/1703.04599.pdf [PDF]).
- In Spring 2017, I am teaching a special topic course for graduate students at our department
- Title: Selected topics in modern convex optimization: theory, algorithms and applications.
- The tentative syllabus is available [HERE].
I am an assistant professor at the Department of Statistics and Operations Research, where I joined in July 2015. I was a postdoctoral researcher at Laboratory for Information and Inference Systems (LIONS), École Polytechnique Fédérale de Lausanne (EPFL), Switzerland from November 2012 to June 2015. I completed my Ph.D. in Optimization in Engineering in November 2012 at the Department of Electrical Engineering (ESAT) and Optimization in Engineering Center (OPTEC) under the supervision of Prof. Moritz Diehl.