A5021831749- Advanced ceramic materials synthesis
- Ferroelectric and Piezoelectric Materials
- Microwave Dielectric Ceramics Synthesis
- Multiferroics and related materials
- Fault Detection and Control Systems
- Advanced materials and composites
- Advanced Control Systems Optimization
- Spectroscopy and Chemometric Analyses
- Mineral Processing and Grinding
- Aluminum Alloys Composites Properties
- Electronic and Structural Properties of Oxides
- High Entropy Alloys Studies
- Analytical Chemistry and Sensors
- Extremum Seeking Control Systems
- Intermetallics and Advanced Alloy Properties
- Nuclear materials and radiation effects
- Advancements in Solid Oxide Fuel Cells
- Metal Alloys Wear and Properties
- Thermal and Kinetic Analysis
- Thermal properties of materials
- Metal and Thin Film Mechanics
- Advanced Data Processing Techniques
- Neural Networks and Applications
- Aluminum Alloy Microstructure Properties
- Advanced Algorithms and Applications
Hangzhou Dianzi University
2024
Tongji University
2023-2024
Chinese Academy of Medical Sciences & Peking Union Medical College
2023
Peking Union Medical College Hospital
2023
Louisiana State University
2017-2022
China Geological Survey
2021
Ocean University of China
2019-2021
Qingdao Institute of Marine Geology
2021
Louisiana State University Agricultural Center
2020
University of Missouri
2018
This paper presents a novel and practical convolutional neural network architecture to implement semantic segmentation for side scan sonar (SSS) image. As widely used sensor marine survey, SSS provides higher-resolution images of the seafloor underwater target. However, large number background pixels in image, imbalance classification remains an issue. What is more, contain undesirable speckle noise intensity inhomogeneity. We define detail training strategy that tackle these three important...
A generic process visualization method is introduced, which visualizes real-time information and correlations among variables on a 2D map using parametric t-SNE. As an unsupervised learning method, it learns the mapping by minimizing Kullback–Leibler divergence between original high-dimensional space latent deep neural network. In practice, observed that t-SNE lacks generalization struggles to visualize unseen operating conditions correctly. this work, two steps improve its capacity are...
We propose a deep learning-based sensor that mitigates the problem of crystal detection in high-density slurries using segmentation model based on RetinaNet framework. The functions single stage contrast to current state-of-the-art learning models such as Mask R-CNN require two stages. It does this by dividing work among three subnetworks parallel with each one solving slightly different problem. While first subnetwork localizes objects image, second predicts their class labels, final task...
In this paper, an adaptive process monitoring method based on the k-nearest neighbor rule (k-NN) is proposed to address issues arising from nonlinearity, insufficient training data, and time-varying behaviors. Instead of recursively updating every measurement for adaptation, a distance-based applied search target prototypes, thus reducing computational load online implementation. Furthermore, fault identification, subspace greedy also introduced formulate complete system. The approach...
Monitoring the operation of a pyrolysis reactor is always challenging due to extremely high‐operating temperature (over 800°C) in fired furnace. To improve current monitoring capability, framework proposed that builds upon thermal photography provide detailed view inside Based on infrared images generated from data provided by cameras, deep learning approach introduced automatically identify tube regions raw images. The pixel‐wise segmentation network named Res50‐UNet, which combines popular...
Underwater fishing nets represent a danger faced by autonomous underwater vehicles (AUVs). To avoid irreparable damage to the AUV caused nets, needs be able identify and locate them autonomously in advance. Whether can successfully depends on accuracy efficiency of detection. In this paper, we propose an object detection multiple receptive field network (MRF-Net), which is used recognize using forward-looking sonar (FLS) images. The proposed architecture center-point-based detector, uses...
Shale of the Middle–Upper Permian Dalong Formation and Gufeng in Lower Yangtze region is characterized by large thickness, high total organic carbon (TOC), wide distribution moderate thermal evolution degree, so it may be next important field shale gas exploration. In order to point out target direction exploration development this region, paper selects Xuanjing area as research object quantitatively describe characteristics pores different scales means scanning electron microscopy (SEM),...
This paper deals with the design and implementation of optimal real-time control strategies for controlling polymerization reactors in free radical processes. A multiobjective optimization problem is first formulated to determine trajectories a range target products. tailor-made module then implemented python weight-average molar mass (Mw) robustly achieve desired polymer distribution (MMD). Multiple case studies are discussed applied force system along targets. Performance controller...
Abstract In this article, multiple reinforcement learning (RL) methods such as value‐based, policy‐based, and actor‐critic algorithms are investigated for typical control tasks found in the chemical industries. Through a critical assessment of these novel techniques, their main advantages highlighted, but also challenges that still need to be resolved discussed. Two batch used benchmarks, namely, production maximization, setpoint control. Using testing environments, direct comparison...