Xiang Yu

ORCID: 0000-0002-0417-0109
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Advanced Neural Network Applications
  • Domain Adaptation and Few-Shot Learning
  • Metaheuristic Optimization Algorithms Research
  • Human Pose and Action Recognition
  • Advanced Multi-Objective Optimization Algorithms
  • Multimodal Machine Learning Applications
  • Neural Networks and Applications
  • Fault Detection and Control Systems
  • Machine Learning and ELM
  • Advanced Algorithms and Applications
  • Electric Vehicles and Infrastructure
  • Face and Expression Recognition
  • Spectroscopy Techniques in Biomedical and Chemical Research
  • Advanced Battery Technologies Research
  • Mineral Processing and Grinding

NEC (United States)
2022-2023

Nanchang Institute of Technology
2018-2022

This article proposes single-objective/multiobjective cat swarm optimization clustering algorithms for data partition. The proposed methods use the to search optimal. position of tightly associates with centers and is updated by two submodes: seeking mode tracing mode. uses simulated annealing strategy update at a probability. Inspired quantum theories, adopts model in whole solution space. First, single-objective method cohesion as objective function, which kernel applied. For considering...

10.1109/tase.2020.2969485 article EN IEEE Transactions on Automation Science and Engineering 2020-01-01

Multi-task learning commonly encounters competition for resources among tasks, specifically when model capac-ity is limited. This challenge motivates models which al-low control over the relative importance of tasks and total compute cost during inference time. In this work, we pro-pose such a controllable multi-task network that dynami-cally adjusts its architecture weights to match de-sired task preference as well resource constraints. contrast existing dynamic approaches adjust only...

10.1109/cvpr52688.2022.01068 article EN 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2022-06-01

Multiswarm comprehensive learning particle swarm optimization (MSCLPSO) is a multiobjective metaheuristic recently proposed by the authors. MSCLPSO uses multiple swarms of particles and externally stores elitists that are nondominated solutions found so far. can approximate true Pareto front in one single run; however, it requires large number generations to converge, because each only optimizes associated objective does not learn from any search experience outside swarm. In this paper, we...

10.3390/info9070173 article EN cc-by Information 2018-07-15

This paper presents an approach to train a unified deep network that simultaneously solves multiple human-related tasks. A multi-task framework is favorable for sharing information across tasks under restricted computational resources. However, not only share but may also compete resources and conflict with each other, making the optimization of shared parameters difficult leading suboptimal performance. We propose simple effective training scheme called GradSplit alleviates this issue by...

10.1109/wacv56688.2023.00433 article EN 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2023-01-01

Battery degradation in electric vehicles is detrimental and keeping track of the State Health (SOH) battery essential. Current SOH estimation techniques (such as monitoring voltage decrease with increasing cycles) have several drawbacks, such having monotonically decreasing predictions non-monotonic systems susceptibility to noise effects. There a simpler more reliable technique by charging times every cycle. This method proven be accurate low Results show nearly identical results scenarios...

10.1109/iccve52871.2022.9743069 article EN 2022-03-07

Ball mill pulverizing system (BMPS) is an important equipment in the thermal power plant and working conditions classification premise of realizing control optimization. This paper proposes ball based on quantum cat swarm optimization clustering algorithm. Firstly, this algorithm creates N cats, representing kinds data classifications respectively, to initialize swarm. The position each coded by centers. And then cats are randomly divided into searching mode tracing find their optimal...

10.1109/yac.2016.7804917 article EN 2016-11-01
Coming Soon ...