A5101447106- Advanced Graph Neural Networks
- Topic Modeling
- Face recognition and analysis
- Robotic Path Planning Algorithms
- Face and Expression Recognition
- Natural Language Processing Techniques
- Biometric Identification and Security
- Access Control and Trust
- Interactive and Immersive Displays
- Robotic Mechanisms and Dynamics
- Robotics and Sensor-Based Localization
- Tissue Engineering and Regenerative Medicine
- Internet Traffic Analysis and Secure E-voting
- Data Quality and Management
- Reproductive Biology and Fertility
- Mesenchymal stem cell research
- Control and Dynamics of Mobile Robots
- Recommender Systems and Techniques
- Domain Adaptation and Few-Shot Learning
- Underwater Vehicles and Communication Systems
- Augmented Reality Applications
- Maritime Navigation and Safety
- Expert finding and Q&A systems
Harbin Engineering University
2024
NetEase (China)
2022
Chinese Academy of Sciences
2020-2021
Capital Normal University
2021
Soochow University
2021
First Affiliated Hospital of Soochow University
2021
Suzhou Institute of Biomedical Engineering and Technology
2021
University of Chinese Academy of Sciences
2018-2020
Xi'an Institute of Optics and Precision Mechanics
2020
Xidian University
2020
This paper proposes a fusion algorithm based on state-tracking collision detection and the simulated annealing potential field (SCD-SAPF) to address challenges of obstacle avoidance for autonomous underwater vehicles (AUVs) in dynamic environments. Navigating AUVs complex environments requires robust capabilities. The SCD-SAPF aims accurately assess risks efficiently plan trajectories. introduces an SCD model proactive risk assessment, predicting between obstacles. Additionally, it (SA)...
In women of reproductive age, severe injuries to the ovary are often accompanied by premature ovarian failure (POF), which can result in amenorrhea or infertility. Hormone replacement therapy has been used treat POF; however, it limited therapeutic efficiency and may cause several side effects. this study, we aimed fabricate a Matrigel scaffold loaded with human umbilical cord-derived mesenchymal stem cells (MSCs) explore its potential restore function repair structures vitro vivo. POF mouse...
Aiming at a series of requirements obstacle avoidance trajectory planning manipulators, new algorithm based on six-order polynomial is proposed. Firstly, the used for manipulator. Assuming that coefficients sixth order term in curve equation are undetermined parameters, by adjusting these shape can be changed to make manipulators avoid and optimize performance indicators simultaneously. Thus, transformed into multi-objective optimization problem. Secondly, combining collision detection...
Cross-domain recommendation (CDR) methods are proposed to tackle the sparsity problem in click through rate (CTR) estimation. Existing CDR directly transfer knowledge from source domains target domain and ignore heterogeneities among domains, including feature dimensional heterogeneity latent space heterogeneity, which may lead negative transfer. Besides, most of existing based on single-source transfer, cannot simultaneously utilize multiple further improve model performance domain. In this...
Augmented reality (AR) has progressed to the point where geometry-aware real-time applications involving multiple users are now possible. We present an approach for a collaborative augmented environment using RGB-D camera and KinectFusion, collecting visual depth data from static that is used as fiducial users. This allows environments digital real world objects can appear interact. anticipate combination of technologies we in our prototype will soon be available mobile devices support reality.
Recent progress of face recognition benefits a lot from large-scale datasets with deep Convoluitonal Neural Networks(CNN). However, when dataset contains large number subjects but few samples for each subject, conventional CNN softmax loss is heavily prone to overfitting. To address this issue, we propose hierarchical training schema optimize coarse-to-fine class labels, referred as Hit-CNN. Firstly trained coarse labels and then refined fine Hit-CNN enabled the capture distribution data...
Theimplicit discourse relation classification is of great importance to analysis. It aims identify the logical between sentence pair. Compared with linear network model, graph neural has a more complex structure capture cross-sentence interactions. Therefore, this article proposes semantic for implicit classification. Specifically, we design describe syntactic sentences and interactions Then, convolutional (CNN) different kernels extract multi-granularity features. The experimental results...