Results 131 to 140 of about 32,087 (238)

Enhancing Generalisation via Cascaded Inertia SGD With Learnt Hyperparameters

open access: yesCAAI Transactions on Intelligence Technology, EarlyView.
ABSTRACT A central challenge in deep learning lies in achieving strong model generalisation, an area in which conventional optimisers such as stochastic gradient descent (SGD) often exhibit limitations, even though they ensure convergence. This paper introduces cascaded inertia SGD (CISGD), a novel optimisation algorithm specifically designed to ...
Yongji Guan   +3 more
wiley   +1 more source

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

open access: yesCAAI 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   +3 more
wiley   +1 more source

Convergent Interval Confirmation for Three‐Step Discrete Specific Zeroing Neural Dynamics Illustrated With Repetitive Motion Planning of Redundant Manipulators

open access: yesCAAI Transactions on Intelligence Technology, EarlyView.
ABSTRACT Repetitive motion planning (RMP) for redundant manipulators with high convergent precision becomes an intense research topic due to its more degrees of freedom. In this paper, a specific zeroing neural dynamics (SZND) model for the RMP is first set up via zeroing neurodynamics.
Ying Kong   +3 more
wiley   +1 more source

A Robust Visual Inertial Odometry SLAM Considering Robot Self Dynamics

open access: yesCAAI Transactions on Intelligence Technology, EarlyView.
ABSTRACT In this paper, to deal with the dynamic SLAM problem, we investigate feature tracking and IMU preintegration in visual‐inertial odometry (VIO) and design a robust SLAM framework that explicitly considers robot self‐dynamics. We propose a self‐dynamics and IMU‐aided feature tracker to predict initial optical flow and an iterative refinement ...
Junyin Qiu, Hong Liu, Tianwei Zhang
wiley   +1 more source

Low‐Frequency Oscillation Suppression of FDR Time‐Domain Diagnostic Waveforms for Cable Defects Based on Splicing Technology

open access: yesHigh Voltage, EarlyView.
ABSTRACT A time domain pulse (TDP) based on the frequency domain reflectometry (FDR) method can be used to diagnose the insulation state of cable defects. However, because of the limitations of the FDR test instrument, the lack of a reflection coefficient spectrum (RCS) in the low‐frequency band (LFB) can result in low frequency oscillation in the time
Jingtao Huang   +6 more
wiley   +1 more source

Home - About - Disclaimer - Privacy