Thomas Pethick

25 papers A* 10Journal 15
YearRankTypeTitle / Venue / Authors
2025 A* conf
Efficient Interpolation between Extragradient and Proximal Methods for Weak MVIs.
ICLR
Thomas Pethick, Ioannis Mavrothalassitis, Volkan Cevher
2025 J jnl
Generalized Gradient Norm Clipping & Non-Euclidean (L
CoRR
Thomas Pethick, Wanyun Xie, Mete Erdogan, Kimon Antonakopoulos, Tony Silveti-Falls, Volkan Cevher
2025 A* conf
Training Deep Learning Models with Norm-Constrained LMOs.
ICML
Thomas Pethick, Wanyun Xie, Kimon Antonakopoulos, Zhenyu Zhu, Antonio Silveti-Falls, Volkan Cevher
2025 J jnl
Training Deep Learning Models with Norm-Constrained LMOs.
CoRR
Thomas Pethick, Wanyun Xie, Kimon Antonakopoulos, Zhenyu Zhu, Antonio Silveti-Falls, Volkan Cevher
2025 J jnl
Training Neural Networks at Any Scale.
CoRR
Thomas Pethick, Kimon Antonakopoulos, Antonio Silveti-Falls, Leena Chennuru Vankadara, Volkan Cevher
2025 J jnl
νSAM: Memory-Efficient Sharpness-Aware Minimization via Nuclear Norm Constraints.
Trans. Mach. Learn. Res.
Thomas Pethick, Parameswaran Raman, Lenon Minorics, Mingyi Hong, Shoham Sabach, Volkan Cevher
2024 A* conf
Improving SAM Requires Rethinking its Optimization Formulation.
ICML
Wanyun Xie, Fabian Latorre, Kimon Antonakopoulos, Thomas Pethick, Volkan Cevher
2024 J jnl
Improving SAM Requires Rethinking its Optimization Formulation.
CoRR
Wanyun Xie, Fabian Latorre, Kimon Antonakopoulos, Thomas Pethick, Volkan Cevher
2024 J jnl
Mixed Nash for Robust Federated Learning.
Trans. Mach. Learn. Res.
Wanyun Xie, Thomas Pethick, Ali Ramezani-Kebrya, Volkan Cevher
2024 A* conf
SAMPa: Sharpness-aware Minimization Parallelized.
NeurIPS
Wanyun Xie, Thomas Pethick, Volkan Cevher
2024 J jnl
SAMPa: Sharpness-aware Minimization Parallelized.
CoRR
Wanyun Xie, Thomas Pethick, Volkan Cevher
2023 J jnl
Escaping limit cycles: Global convergence for constrained nonconvex-nonconcave minimax problems.
CoRR
Thomas Pethick, Puya Latafat, Panagiotis Patrinos, Olivier Fercoq, Volkan Cevher
2023 J jnl
Federated Learning under Covariate Shifts with Generalization Guarantees.
CoRR
Ali Ramezani-Kebrya, Fanghui Liu, Thomas Pethick, Grigorios Chrysos, Volkan Cevher
2023 J jnl
Federated Learning under Covariate Shifts with Generalization Guarantees.
Trans. Mach. Learn. Res.
Ali Ramezani-Kebrya, Fanghui Liu, Thomas Pethick, Grigorios Chrysos, Volkan Cevher
2023 A* conf
Finding Actual Descent Directions for Adversarial Training.
ICLR
Fabian Latorre, Igor Krawczuk, Leello Tadesse Dadi, Thomas Pethick, Volkan Cevher
2023 J jnl
Revisiting adversarial training for the worst-performing class.
CoRR
Thomas Pethick, Grigorios G. Chrysos, Volkan Cevher
2023 J jnl
Revisiting adversarial training for the worst-performing class.
Trans. Mach. Learn. Res.
Thomas Pethick, Grigorios Chrysos, Volkan Cevher
2023 A* conf
Solving stochastic weak Minty variational inequalities without increasing batch size.
ICLR
Thomas Pethick, Olivier Fercoq, Puya Latafat, Panagiotis Patrinos, Volkan Cevher
2023 J jnl
Solving stochastic weak Minty variational inequalities without increasing batch size.
CoRR
Thomas Pethick, Olivier Fercoq, Puya Latafat, Panagiotis Patrinos, Volkan Cevher
2023 A* conf
Stable Nonconvex-Nonconcave Training via Linear Interpolation.
NeurIPS
Thomas Pethick, Wanyun Xie, Volkan Cevher
2023 J jnl
Stable Nonconvex-Nonconcave Training via Linear Interpolation.
CoRR
Thomas Pethick, Wanyun Xie, Volkan Cevher
2022 A* conf
Escaping limit cycles: Global convergence for constrained nonconvex-nonconcave minimax problems.
ICLR
Thomas Pethick, Puya Latafat, Panos Patrinos, Olivier Fercoq, Volkan Cevher
2021 A* conf
Sifting through the noise: Universal first-order methods for stochastic variational inequalities.
NeurIPS
Kimon Antonakopoulos, Thomas Pethick, Ali Kavis, Panayotis Mertikopoulos, Volkan Cevher
2021 A* conf
Subquadratic Overparameterization for Shallow Neural Networks.
NeurIPS
Chaehwan Song, Ali Ramezani-Kebrya, Thomas Pethick, Armin Eftekhari, Volkan Cevher
2021 J jnl
Subquadratic Overparameterization for Shallow Neural Networks.
CoRR
Chaehwan Song, Ali Ramezani-Kebrya, Thomas Pethick, Armin Eftekhari, Volkan Cevher