Results 71 to 80 of about 6,455,829 (284)
Invariance properties of the natural gradient in overparametrised systems
, 2023 The natural gradient field is a vector field that lives on a model equipped with a distinguished Riemannian metric, e.g. the Fisher–Rao metric, and represents the direction of steepest ascent of an objective function on the model with respect to this ...
Müller, Johannes +2 more
core +1 more source
Improving Flat Maxima with Natural Gradient for Better Adversarial Transferability
Big Data and Cognitive ComputingDeep neural networks are vulnerable and susceptible to adversarial examples, which can induce erroneous predictions by injecting imperceptible perturbations.
Yunfei Long, Huosheng Xu
doaj +1 more source
Natural Gradient Descent for On-Line Learning
Physical Review Letters, 1998 Natural gradient descent is an on-line variable-metric optimization algorithm which utilizes an underlying Riemannian parameter space. We analyze the dynamics of natural gradient descent beyond the asymptotic regime by employing an exact statistical mechanics description of learning in two-layer feed-forward neural networks.
Magnus Rattray +2 more
openaire +2 more sources
How structural disorder and polyelectrolyte properties of prothymosin α enable its large interactome
FEBS Letters, EarlyView.In this review, complexes of the fully disordered and highly negatively charged prothymosin α are collected, described, and discussed. Linked to its poly‐electrostatic character, prothymosin α engages a large interactome through multiple different binding modes ranging from fully disordered complexes to short linear motif (SLiM)‐driven interactions ...
Ketty C. Tamburrini +3 more
wiley +1 more source
Properties of functionally gradient composites reinforced with waste natural fillers [PDF]
Acta Periodica Technologica, 2019 In the present work, an in-house centrifugal casting setup was developed in the laboratory. Waste natural fillers such as soap nuts seeds, Aegle marmelos, and Terminalia chebula were used for the preparation of polyester based composite by using
Prasad Lalta +4 more
doaj +1 more source
Natural gradient descent with momentum
CoRRWe consider the problem of approximating a function by an element of a nonlinear manifold which admits a differentiable parametrization, typical examples being neural networks with differentiable activation functions or tensor networks. Natural gradient descent (NGD) for the optimization of a loss function can be seen as a preconditioned gradient ...
Nouy, Anthony, Somacal, Agustín
openaire +4 more sources
Aquaporin‐3 and aquaporin‐5 impact the development of pancreatic ductal adenocarcinoma spheroids
FEBS Open Bio, EarlyView.Schematic representation of the role of aquaporin‐3 (AQP3) and aquaporin‐5 (AQP5) in pancreatic ductal adenocarcinoma (PDAC). Both proteins are upregulated in PDAC and are associated with tumor progression and metastatic potential. Silencing AQP3 or AQP5 in PDAC spheroids results in decreased diameter, area, and overall growth, underscoring their key ...
Catarina Pimpão +3 more
wiley +1 more source
Enhanced gradient learning for deep neural networks
IET Image Processing, 2022 Deep neural networks have achieved great success in both computer vision and natural language processing tasks. How to improve the gradient flows is crucial in training very deep neural networks. To address this challenge, a gradient enhancement approach
Ming Yan +5 more
doaj +1 more source
Directed evolution of enzymes at the crossroads of tradition and innovation
FEBS Open Bio, EarlyView.An iterative cycle of data‐driven enzyme optimization comprising four stages: genetic diversification of a template enzyme, expression of protein variants, high‐throughput evaluation, and machine‐learning‐guided redesign of the next variant library.
Maria Tomkova +2 more
wiley +1 more source
Variable-step blind source separation method with adaptive momentum factor
Tongxin xuebao, 2017 A variable-step blind source separation algorithm based on the natural gradient with adaptive momentum factor was proposed,which could cope with the determined blind source separation in the environment of stationary and non-stationary.Function ...
Tian-qi ZHANG +3 more
doaj