Results 131 to 140 of about 36,144 (280)

SGD-YOLOv5: A Small Object Detection Model for Complex Industrial Environments

SGD-YOLOv5: A Small Object Detection Model for Complex Industrial ...
J Pei (13464954), X Wu (660808), S Yu (7291433), X Zheng (7940567), X Liu (7756691), Longxiang Gao (13081947)   +5 more
core  

Respiratory Variability during Exercise: Associations with Motor Severity in Parkinson's Disease

Movement Disorders Clinical Practice, EarlyView.
Marcela S. Araújo, Muhammad M. Kathia, Sergiu‐Gabriel Duplea, Julian C. Bommarito, Devin G. McCarthy, Barbara Connolly, Philip J. Millar, Lauro C. Vianna   +7 more
wiley   +1 more source

Comparative Genomics and Heterologous Complementation Reveal Sequence and Structural Divergence of Selected Transcription Factors From Spathaspora passalidarum

Yeast, EarlyView.
Comparative genomic analyses identified substantial sequence and structural differences between transcription factors from Spathaspora passalidarum and their Saccharomyces cerevisiae counterparts. Heterologous expression of YAP1 and MSN‐like genes in corresponding S.
Fernanda Matias Albuini, Alex Gazolla de Castro, Lílian Emídio Ribeiro, Lizandra Cristina de Oliveira Figueiredo Gazolla, Carlos Augusto Rosa, Tiago Antônio de Oliveira Mendes, Luciano Gomes Fietto   +6 more
wiley   +1 more source

Adaptive Step-Size Methods for Compressed SGD

, 2022
Compressed Stochastic Gradient Descent (SGD) algorithms have been recently proposed to address the communication bottleneck in distributed and decentralized optimization problems, such as those that arise in federated machine learning.
Subramaniam, Adarsh M., Veeravalli, Venugopal V., Magesh, Akshayaa   +2 more
core  

Shaoyao Gancao Decoction: a comprehensive review of modern clinical applications and underlying pharmacological mechanisms

Frontiers in Pharmacology
Shaoyao Gancao Decoction, a classic formula from Zhang Zhongjing in the Han Dynasty, has the effects of regulating the function of the liver, relieving spasm and stopping pain.
Du Hong, Ying Xiao
doaj   +1 more source

Generalization and Optimization of SGD with Lookahead

CoRR
The Lookahead optimizer enhances deep learning models by employing a dual-weight update mechanism, which has been shown to improve the performance of underlying optimizers such as SGD. However, most theoretical studies focus on its convergence on training data, leaving its generalization capabilities less understood.
Kangcheng Li, Yunwen Lei
openaire   +3 more sources

MAPFL: Modular Adaptive Personalised Federated Learning From Visual Sentiment to Physical Activity

CAAI Transactions on Intelligence Technology, EarlyView.
ABSTRACT Human‐centric sensing tasks such as visual sentiment analysis and smartphone‐based human activity recognition (HAR) exhibit strong client heterogeneity from feature, label, and concept shifts. Among these, concept shift—how individual users interpret and generate human‐centric signals—is particularly challenging yet underexplored in federated ...
Chang Liu, Yapeng Li, Zengmao Wang, Dewen Soh, Bo Du   +4 more
wiley   +1 more source

Double‐Integration‐Enhanced Stochastic Gradient Descent Based on Neural Dynamics for Improving Generalisation

CAAI Transactions on Intelligence Technology, EarlyView.
ABSTRACT Generalisation is a crucial aspect of deep learning, enabling models to perform well on unseen data. Currently, most optimisers that improve generalisation typically suffer from efficiency bottlenecks. This paper proposes a double‐integration‐enhanced stochastic gradient descent (DIESGD) optimiser, which treats the negative gradient as an ...
Ting Li, Longbang Wang, Jiexing Li, Jicheng Li   +3 more
wiley   +1 more source

Robust Multi‐Source Batch Normalisation for Test‐Time Batch Adaptation

CAAI Transactions on Intelligence Technology, EarlyView.
ABSTRACT Test‐Time Batch Adaptation (TTBA) aims to adapt a pre‐trained source model to small, unlabelled target batches at test time. The TTBA methods focus on adapting the pre‐trained model or the target data in a one‐to‐one alignment paradigm. However, these one‐to‐one alignment paradigms assume that the source domain may share the same knowledge ...
Xinlin Xiao, Xiangtao Zheng, Jianhua Tao, Xiaoqiang Lu   +3 more
wiley   +1 more source

On the Trajectories of SGD Without Replacement [PDF]

This article examines the implicit regularization effect of Stochastic Gradient Descent (SGD). We consider the case of SGD without replacement, the variant typically used to optimize large-scale neural networks.
Beneventano, Pierfrancesco
core   +1 more source