Results 11 to 20 of about 836,502 (333)
A complete and operational resource theory of measurement sharpness [PDF]
Quantum, 2023 We construct a resource theory of $sharpness$ for finite-dimensional positive operator-valued measures (POVMs), where the $sharpness-non-increasing$ operations are given by quantum preprocessing channels and convex mixtures with POVMs whose elements are ...
Francesco Buscemi +2 more
doaj +3 more sources
Generalized Federated Learning via Sharpness Aware Minimization [PDF]
International Conference on Machine Learning, 2022 Federated Learning (FL) is a promising framework for performing privacy-preserving, distributed learning with a set of clients. However, the data distribution among clients often exhibits non-IID, i.e., distribution shift, which makes efficient ...
Zhe Qu +5 more
semanticscholar +1 more source
Towards Understanding Sharpness-Aware Minimization [PDF]
International Conference on Machine Learning, 2022 Sharpness-Aware Minimization (SAM) is a recent training method that relies on worst-case weight perturbations which significantly improves generalization in various settings.
Maksym Andriushchenko +1 more
semanticscholar +1 more source
Sharpness-Aware Gradient Matching for Domain Generalization [PDF]
Computer Vision and Pattern Recognition, 2023 The goal of domain generalization (DG) is to enhance the generalization capability of the model learned from a source domain to other unseen domains. The recently developed Sharpness-Aware Minimization (SAM) method aims to achieve this goal by minimizing
Pengfei Wang +3 more
semanticscholar +1 more source
Surrogate Gap Minimization Improves Sharpness-Aware Training [PDF]
International Conference on Learning Representations, 2022 The recently proposed Sharpness-Aware Minimization (SAM) improves generalization by minimizing a \textit{perturbed loss} defined as the maximum loss within a neighborhood in the parameter space. However, we show that both sharp and flat minima can have a
Juntang Zhuang +8 more
semanticscholar +1 more source
Enhancing Fine-Tuning based Backdoor Defense with Sharpness-Aware Minimization [PDF]
IEEE International Conference on Computer Vision, 2023 Backdoor defense, which aims to detect or mitigate the effect of malicious triggers introduced by attackers, is becoming increasingly critical for machine learning security and integrity. Fine-tuning based on benign data is a natural defense to erase the
Mingli Zhu +4 more
semanticscholar +1 more source
A modern look at the relationship between sharpness and generalization [PDF]
International Conference on Machine Learning, 2023 Sharpness of minima is a promising quantity that can correlate with generalization in deep networks and, when optimized during training, can improve generalization. However, standard sharpness is not invariant under reparametrizations of neural networks,
Maksym Andriushchenko +4 more
semanticscholar +1 more source
Towards Efficient and Scalable Sharpness-Aware Minimization [PDF]
Computer Vision and Pattern Recognition, 2022 Recently, Sharpness-Aware Minimization (SAM), which connects the geometry of the loss landscape and generalization, has demonstrated a significant performance boost on training large-scale models such as vision transformers.
Y. Liu +4 more
semanticscholar +1 more source
An Adaptive Policy to Employ Sharpness-Aware Minimization [PDF]
International Conference on Learning Representations, 2023 Sharpness-aware minimization (SAM), which searches for flat minima by min-max optimization, has been shown to be useful in improving model generalization.
Weisen Jiang +3 more
semanticscholar +1 more source
Sharpness-Aware Training for Free [PDF]
Neural Information Processing Systems, 2022 Modern deep neural networks (DNNs) have achieved state-of-the-art performances but are typically over-parameterized. The over-parameterization may result in undesirably large generalization error in the absence of other customized training strategies ...
Jiawei Du +4 more
semanticscholar +1 more source