A5102859112- Data Visualization and Analytics
- Domain Adaptation and Few-Shot Learning
- Complex Network Analysis Techniques
- Cancer-related molecular mechanisms research
- Machine Learning and Data Classification
- Generative Adversarial Networks and Image Synthesis
- Multimodal Machine Learning Applications
- Computational Physics and Python Applications
- Water Quality Monitoring Technologies
- Vehicle emissions and performance
- Anomaly Detection Techniques and Applications
- Internet Traffic Analysis and Secure E-voting
- Privacy-Preserving Technologies in Data
- Time Series Analysis and Forecasting
- Forensic Entomology and Diptera Studies
- Advanced Vision and Imaging
- Forensic Anthropology and Bioarchaeology Studies
- Electrical Contact Performance and Analysis
- Explainable Artificial Intelligence (XAI)
- Data Analysis with R
- Data Mining Algorithms and Applications
- Adversarial Robustness in Machine Learning
- Railway Systems and Energy Efficiency
- Railway Engineering and Dynamics
- Membrane Separation Technologies
Northeast Forestry University
2025
Zhejiang University
2016-2024
Tiangong University
2023
Nankai University
2023
China Agricultural University
2023
Soochow University
2023
Ministry of Agriculture and Rural Affairs
2023
Southwest University
2023
Institute of Forensic Science
2020
Jimei University
2010-2013
The approach for estimating biomass in non-contact, free-swimming fish has encountered difficulties such as body occlusion, bending, non-orthogonal angle, and low efficiency. To address these issues, this study had combined posture recognition (using deep learning technology) with estimation (utilizing stereo vision the first time, developing a fast, precise fully automatic system. improved single-stage target detection algorithm significantly direct extraction of high-quality images moving...
Traditional tumor treatments have certain limitations which serious negative effects, such as the difficulty in cutting out tumors, damage to normal tissues, and complications. Ordinary nanozymes generally low catalytic activity require higher doses for treatment, can increase toxicity side effects. Here, we designed a Cu(2‐x)Se‐loaded on MXene, catalyze conversion of overexpressed hydrogen peroxide into O2 ‐OH cancer cells by Cu(2‐x)Se with catalase (CAT)‐like peroxidase (POD) activity,...
Conventional unsupervised multi-source domain adaptation (UMDA) methods assume all source domains can be accessed directly. This neglects the privacy-preserving policy, that is, data and computations must kept decentralized. There exists three problems in this scenario: (1) Minimizing distance requires pairwise calculation of from target domains, which is not accessible. (2) The communication cost privacy security limit application UMDA (e.g., adversarial training). (3) Since users have no...
Semi-supervised variational autoencoders (VAEs) have obtained strong results, but also encountered the challenge that good ELBO values do not always imply accurate inference results.In this paper, we investigate and propose two causes of problem: (1) The objective cannot utilize label information directly. (2) A bottleneck value exists, continuing to optimize after will improve accuracy. On basis experiment SHOT-VAE address these problems without introducing additional prior knowledge....
The past decade has witnessed rapid progress in AI research since the breakthrough deep learning. technology been applied almost every field; therefore, technical and non-technical end-users must understand these technologies to exploit them. However existing materials are designed for experts, but users need appealing that deliver complex ideas easy-to-follow steps. One notable tool fits such a profile is scrollytelling, an approach storytelling provides readers with natural rich experience...
The development of digitized humanity information provides a new perspective on data-oriented studies history. Many previous have ignored uncertainty in the exploration historical figures and events, which has limited capability researchers to capture complex processes associated with phenomena. We propose visual reasoning system support spatio-temporal events based data from China Biographical Database Project. build knowledge graph entities extracted database generated by missing error....
The widespread use of chemicals has brought serious water pollution threatening human health and environment, which requires green, fast, low-cost purification urgently. Here, we build up a novel material family sky-parking-like 3D structured graphene oxides (SP-GOs) with adjustable interlayer-space 0.8-1.7 nm via the insertion different sized diamine compounds as support pillars between GO layers. assembled SP-GOs exhibit superior adsorption capacity short removal time for various aromatic...
Patterns in charts refer to interesting visual features or forms. Identifying patterns not only helps analysts understand the ‘shape’ of data but also supports better and faster decision-making. Existing solutions for identifying require a large number labeled instances, making it intractable without user supervision. In this paper, we propose ChartNavigator, an interactive pattern identification annotation framework unlabeled visualization charts. ChartNavigator leverages novel...
The study introduced the general evaluation indicator system for community health services in Shanghai and its characteristics, analyzing results of comprehensive from aspects regions institutions. From six financial input, human resource construction, operation mechanism, family doctor system, information construction application results, paper recommended on deepening reform services. Key words: Community services; Performance evaluation; Indicator system; Family
To hold the developing trend of virtual reality technology, and reflect major nations' advances influences on state art, articles indexed by SCI-EXPANDED from 1997–2009 were studied using bibliometric analysis. The only keyword is reality, 2921 analyzed. Some results had been presented in this paper. First, amount papers about world steady growth within last decade, contributed China significant recent years; second, computer science medicine are main subject areas reality; one that citation...
This paper presents GraphFederator, a novel approach to construct federated representations of multi-party graphs and supports privacy-preserving visual analysis graphs. Inspired by the concept learning, we reformulate into decentralization process. The new federation framework consists shared module that is responsible for modeling analysis, set local modules run on respective graph data. Specifically, propose Federated Graph Representation Model (FGRM) learned from encrypted...
Semi-supervised variational autoencoders (VAEs) have obtained strong results, but also encountered the challenge that good ELBO values do not always imply accurate inference results. In this paper, we investigate and propose two causes of problem: (1) The objective cannot utilize label information directly. (2) A bottleneck value exists continuing to optimize after will improve accuracy. On basis experiment SHOT-VAE address these problems without introducing additional prior knowledge....
Insights in power grid pixel maps (PGPMs) refer to important facility operating states and unexpected changes the grid. Identifying insights helps analysts understand collaboration of various parts so that preventive correct operations can be taken avoid potential accidents. Existing solutions for identifying PGPMs are performed manually, which may laborious expertise-dependent. In this paper, we propose an interactive insight identification annotation framework by leveraging enhanced...