Zobrazeno 1 - 10
of 36
pro vyhledávání: '"Chavdarova, Tatjana"'
Multi-agent reinforcement learning (MARL) presents unique challenges as agents learn strategies through experiences. Gradient-based methods are often sensitive to hyperparameter selection and initial random seed variations. Concurrently, significant
Externí odkaz:
http://arxiv.org/abs/2410.07976
Publikováno v:
ICLR 2024
Yang et al. (2023) recently showed how to use first-order gradient methods to solve general variational inequalities (VIs) under a limiting assumption that analytic solutions of specific subproblems are available. In this paper, we circumvent this as
Externí odkaz:
http://arxiv.org/abs/2210.15659
Algorithms that solve zero-sum games, multi-objective agent objectives, or, more generally, variational inequality (VI) problems are notoriously unstable on general problems. Owing to the increasing need for solving such problems in machine learning,
Externí odkaz:
http://arxiv.org/abs/2207.07105
Publikováno v:
International Conference on Learning Representations 2023, Kigali, Rwanda
We develop an interior-point approach to solve constrained variational inequality (cVI) problems. Inspired by the efficacy of the alternating direction method of multipliers (ADMM) method in the single-objective context, we generalize ADMM to derive
Externí odkaz:
http://arxiv.org/abs/2206.10575
Autor:
Pagliardini, Matteo, Manunza, Gilberto, Jaggi, Martin, Jordan, Michael I., Chavdarova, Tatjana
Recently Shah et al., 2020 pointed out the pitfalls of the simplicity bias - the tendency of gradient-based algorithms to learn simple models - which include the model's high sensitivity to small input perturbations, as well as sub-optimal margins. I
Externí odkaz:
http://arxiv.org/abs/2202.05737
Publikováno v:
Minimax Theory 8, Number 2 (2023) 333--380
Several widely-used first-order saddle-point optimization methods yield an identical continuous-time ordinary differential equation (ODE) that is identical to that of the Gradient Descent Ascent (GDA) method when derived naively. However, the converg
Externí odkaz:
http://arxiv.org/abs/2112.13826
Uncertainty estimation (UE) techniques -- such as the Gaussian process (GP), Bayesian neural networks (BNN), Monte Carlo dropout (MCDropout) -- aim to improve the interpretability of machine learning models by assigning an estimated uncertainty value
Externí odkaz:
http://arxiv.org/abs/2112.05000
Data augmentation is a widely adopted technique for avoiding overfitting when training deep neural networks. However, this approach requires domain-specific knowledge and is often limited to a fixed set of hard-coded transformations. Recently, severa
Externí odkaz:
http://arxiv.org/abs/2108.07958
This report is an account of the authors' experiences as organizers of WiML's "Un-Workshop" event at ICML 2020. Un-workshops focus on participant-driven structured discussions on a pre-selected topic. For clarity, this event was different from the "W
Externí odkaz:
http://arxiv.org/abs/2012.01191
Autor:
Chavdarova, Tatjana, Pagliardini, Matteo, Stich, Sebastian U., Fleuret, Francois, Jaggi, Martin
Publikováno v:
ICLR 2021
Generative Adversarial Networks are notoriously challenging to train. The underlying minmax optimization is highly susceptible to the variance of the stochastic gradient and the rotational component of the associated game vector field. To tackle thes
Externí odkaz:
http://arxiv.org/abs/2006.14567