Publications

Journals

1. Adaptive Bilevel Optimization

K. Antonakopoulos; S. Sabach; L. Viano; M. Hong; V. Cevher ACM / IMS Journal of Data Science. 2025. DOI : 10.1145/3728478.

2. On the Generalization of Stochastic Gradient Descent with Momentum 

A. Ramezani-Kebrya; K. Antonakopoulos; V. Cevher; A. Khisti; B. Liang Journal of Machine Learning Research. 2024. Vol. 25, p. 1 – 56.

Conference papers

1. Generalized Gradient Norm Clipping & Non-Euclidean (L_0,L_1)- Smoothness

T. Pethick; W. Xie; K. Antonakopoulos; Mete Erdogan; A. Silveti-Falls et al. 2025. Forty-Second International Conference on Machine Learning, Vancouver, Canada, 2025-07-13 – 2025-07-19. (Oral presentation)

2. Layer-wise Quantization for Quantized Optimistic Dual Averaging

A. Duc Nguyen; I. Markov; F. Z. Wu; A. Ramezani-Kebrya; K. Antonakopoulos et al. 2025. Forty-Second International Conference on Machine Learning, Vancouver, Canada, 2025-07-13 – 2025-07-19.

3. Training Deep Learning Models with Norm-Constrained LMOs

T. Pethick; W. Xie; K. Antonakopoulos; Z. Zhu; A. Silveti-Falls et al. 2025. Forty-Second International Conference on Machine Learning, Vancouver, Canada, 2025-07-13 – 2025-07-19. (Spotlight presentation)

4. Universal Gradient Methods for Stochastic Convex Optimization

A. Rodomanov; A. Kavis; Y. Wu; K. Antonakopoulos; V. Cevher 2024. 41st International Conference on Machine Learning (ICML 2024), Vienna, Austria, 2024-07-21.

5. Improving SAM Requires Rethinking its Optimization Formulation

W. Xie; F. Latorre; K. Antonakopoulos; T. M. Pethick; V. Cevher 2024. 41st International Conference on Machine Learning (ICML 2024), Vienna, Austria, July 21-27, 2024.

6. Advancing the Lower Bounds: An Accelerated, Stochastic, Second-Order Method with Optimal Adaptation to Inexactness

Artem Agafonov; Dmitry Kamzolov; Alexander Gasnikov; Ali Kavis; Kimon Antonakopoulos; Volkan Cevher; Martin Takác

12th International Conference on Learning Representations (ICLR),Vienna Austria, May 7-11, 2024.

7. Distributed Extra-Gradient With Optimal Complexity And Communication Guarantees

A. Ramezani-Kebrya; K. Antonakopoulos; I. Krawczuk; J. Deschenaux; V. Cevher 2023. 11th International Conference on Learning Representations (ICLR), Kigali, Rwanda, May 1-5, 2023.

8. No-regret learning in games with noisy feedback: Faster rates and adaptivity via learning rate separation

Y-G. Hsieh; K. Antonakopoulos; V. Cevher; P. Mertikopoulos 2022. 36th Conference on Neural Information Processing Systems (NeurIPS 2022), New Orleans, USA, November 28 – December 9, 2022.

9. Adaptive Stochastic Variance Reduction for Non-convex Finite-Sum Minimization

A. Kavis; E. P. Skoulakis; K. Antonakopoulos; L. T. Dadi; V. Cevher 2022. 36th Conference on Neural Information Processing Systems (NeurIPS), New Orleans, Louisianna, USA, November 28-December 9, 2022.

10. ADAGRAD Avoids Saddle Points

K. Antonakopoulos; P. Mertikopoulos; G. Piliouras; X. Wang 2022. 38th International Conference on Machine Learning (ICML), Baltimore, MD, Jul 17-23, 2022. p. 731 – 771.

11. Extra Newton: A First Approach to Noise-Adaptive Accelerated Second-Order Methods

K. Antonakopoulos; A. Kavis; V. Cevher 2022. 36th Conference on Neural Information Processing Systems (NeurIPS), New Orleans, Louisianna, USA, November 28-December 9, 2022.

12. UNDERGRAD: A Universal Black-Box Optimization Method with Almost Dimension-Free Convergence Rate Guarantees

K. Antonakopoulos; D. Q. Vu; V. Cevher; K. Y. Levy 2022. 39th International Conference on Machine Learning (ICML), Baltimore, Maryland, USA, July 17-23, 2022. (Spotlight presentation)

13. Sifting through the Noise: Universal First-Order Methods for Stochastic Variational Inequalities

K. Antonakopoulos; T. M. Pethick; A. Kavis; P. Mertikopoulos; V. Cevher 2021. NeurIPS 2021 : Thirty-fifth Conference on Neural Information Processing Systems, Sydney, Australia [Virtual only], December 6-14, 2021.

14. Adaptive First-Order Methods Revisited: Convex Minimization without Lipschitz Requirements

K. Antonakopoulos; P. Mertikopoulos. NeurIPS 2021 : Thirty-fifth Conference on Neural Information Processing Systems, Sydney, Australia [Virtual only], December 6-14, 2021.

15. Fast Routing under Uncertainty: Adaptive Learning in Congestion Games via Exponential Weights

D.-Q. Vu ;K. Antonakopoulos; P. Mertikopoulos. NeurIPS 2021: Thirty-fifth Conference on Neural Information Processing Systems, Sydney, Australia [Virtual only], December 6-14, 2021.

16. Adaptive Learning in Continuous Games: Optimal Regret Bounds and Convergence to Equilibrium

Y-G. Hsieh; K. Antonakopoulos; P. Mertikopoulos. COLT 2021: Proceedings of 34th Annual Conference on Learning Theory, Boulder, Colorado, USA, August 15-21 2021.

17. Adaptive Extra-Gradient Methods for Min-Max Optimization and Games

K. Antonakopoulos; E.V. Belmega; P. Mertikopoulos. 9th International Conference on Learning Representations (ICLR), [Virtual Only]  May 3-7 2021.

18. Online and Stochastic Optimization Beyond Lipschitz Continuity: A Riemannian Approach

K. Antonakopoulos; E.V. Belmega; P. Mertikopoulos. 8th International Conference on Learning Representations (ICLR), [Virtual Only] 26 April-1 May 2020. (Spotlight presentation)

19. An Adaptive Mirror-Prox Method for Variational Inequalities with Singular Operators

K. Antonakopoulos; E.V. Belmega; P. Mertikopoulos. NeurIPS 20219 Proceedings of the 33rd International Conference on Neural Processing Information Systems, 2019.

Preprints

1. Training Neural Networks at Any Scale 

T. Pethick; K. Antonakopoulos; A. Silveti-Falls; L. Chennuru Vankadara; V. Cevher (under minor revision for IEEE Signal Processing Magazine).

2. Multi-Step Alignment as Markov Games: An Optimistic Online Gradient Descent Approach with Convergence Guarantees

Y. Wu; L Viano; Y. Chen; Z. Zhu; K. Antonakopoulos; Q. Gu; V. Cevher (under review for Transactions of Machine Learning Research).

3. Routing in an Uncertain World: Adaptivity, Efficiency and Equilibrium

D.-Q. Vu; K. Antonakopoulos; P. Mertikopoulos.

4. Extra-Gradient Converges in Iterates Faster than O(1/√T ) for Both Constant and Adaptive Step Sizes in All Monotone Lipschitz Variational Inqualities

K. Antonakopoulos; G. Farina; V. Cevher. NeurIPS/OPT2024: 16th Annual Workshop on Optimization for Machine Learning.

5. Bregman Proximal Methods for Stochastic Variational Inequalities with Singular operators

K. Antonakopoulos and P. Mertikopoulos (in preparation).