A5100644758- Advanced Graph Neural Networks
- Recommender Systems and Techniques
- Topic Modeling
- Advanced Vision and Imaging
- Cloud Computing and Resource Management
- Machine Learning in Healthcare
- Lignin and Wood Chemistry
- Robotics and Sensor-Based Localization
- Optical measurement and interference techniques
- Advanced Neural Network Applications
- Sentiment Analysis and Opinion Mining
- Advanced Text Analysis Techniques
- Mental Health via Writing
- Graph Theory and Algorithms
- Corrosion Behavior and Inhibition
- Stock Market Forecasting Methods
- Web Data Mining and Analysis
- Data Management and Algorithms
- Hydrogen embrittlement and corrosion behaviors in metals
- Perovskite Materials and Applications
- Enzyme-mediated dye degradation
- Microfluidic and Capillary Electrophoresis Applications
- Face recognition and analysis
- Financial Markets and Investment Strategies
- Parallel Computing and Optimization Techniques
Chinese Academy of Sciences
2017-2025
China University of Petroleum, East China
2025
Northwest University
2018-2025
Fujian Normal University
2025
Fujian Institute of Research on the Structure of Matter
2023-2025
Northwestern Polytechnical University
2025
Tiangong University
2024
Dongbei University of Finance and Economics
2024
Walmart (United States)
2022-2024
Institute of High Energy Physics
2024
Feature Pyramid Networks (FPN) is a popular feature extraction. However, FPN and its variants do not investigate the influence of resolution information semantic in object detection. Thus, cannot detect some objects on challenging images. In this paper, based FPN, we propose to use gaussian kernel function assign different weight values for images The proposed method, called Weighted Network (WFPN), shows significant improvement over traditional pyramids several applications. Using WFPN...
Dynamic recommendation is essential for modern recommender systems to provide real-time predictions based on sequential data. In real-world scenarios, the popularity of items and interests users change over time. Based this assumption, many previous works focus interaction sequences learn evolutionary embeddings items. However, we argue that sequence-based models are not able capture collaborative information among directly. Here propose Graph Collaborative Filtering (DGCF), a novel...
Mining Electronic Health Records (EHRs) becomes a promising topic because of the rich information they contain. By learning from EHRs, machine models can be built to help human expert make medical decisions and thus improve healthcare quality. Recently, many based on sequential or graph model are proposed achieve this goal. EHRs contain multiple entities relations, viewed as heterogeneous graph. However, previous studies ignore heterogeneity in EHRs. On other hand, current neural networks...
Recently, the percentage of people with hypertension is increasing, and this phenomenon widely concerned. At same time, wireless home Blood Pressure (BP) monitors become accessible in people's life. Since machine learning methods have made important contributions different fields, many researchers tried to employ them dealing medical problems. However, existing studies for BP prediction are all based on clinical data short time ranges. Besides, there do not exist works which can jointly make...
The problem of basket recommendation (BR) is to recommend a ranking list items the current basket. Existing methods solve this by assuming within same are correlated one semantic relation, thus optimizing item embeddings. However, assumption breaks when there exist multiple intents For example, contains {bread, cereal, yogurt, soap, detergent} where yogurt} through "breakfast" intent, while {soap, "cleaning" ignoring relations among spoils ability model learn To resolve issue, it required...
Interneurons are fundamental cells for maintaining the excitation-inhibition balance in brain health and disease. While interneurons have been shown to play a key role pathophysiology of autism spectrum disorder (ASD) adult mice, little is known about how their maturation altered developing striatum ASD. Here, we aimed track striatal elucidate molecular physiological alterations Cntnap2 knockout mouse model. Using Stereo-seq single-cell RNA sequencing data, first characterized pattern...
The electrochemical reduction of CO2 (CO2RR) to value-added chemicals represents a critical strategy for mitigating carbon emissions and promoting energy sustainability. This study focuses on enhancing the performance copper-based catalysts through silver doping, with specific objective improving C2+ product selectivity suppressing C1 products. We report delicate synthesis three distinct CuAg Janus nanostructures using coreduction method involving metal precursors nucleation growth. Compared...
Estimation of the subsurface chlorophyll maximum (SCM) depth is critical to constructing vertical profile a, which in turn important accurately assessing phytoplankton distribution, especially extensive tropical oceans. Current ocean colour algorithms generally detect a limited area only or use suite environmental variables, part may be difficultly quantified. By using field situ observations, two-step approach, first step used relationship between sea surface temperature and occurrence SCM...
To address the inherent scale ambiguity and positioning drift in monocular visual Simultaneous Localization Mapping (SLAM), this paper proposes a novel localization method that integrates SLAM with Ultra-Wideband (UWB) ranging information. This enables high-precision for unmanned aerial vehicles (UAVs) complex environments without global navigation The proposed framework, IVU-AutoNav, relies solely on distance measurements between fixed UWB anchor UAV’s device. Initially, it jointly solves...
Abstract Thermally activated delayed fluorescence (TADF) materials hold great promise as next‐generation efficient emitters due to their theoretical 100% internal quantum efficiency. However, such an intriguing photophysical mechanism is extremely rare in perovskites or derivatives. Herein, the colloidal synthesis of two‐dimensional CsAgCl 2 nanocrystals (NCs) with typical TADF feature reported first. The investigated detail by combining first‐principles calculations experimental results,...
Abstract This paper presents a single-shot phase extraction approach based on deep convolutional generative adversarial network that generates map and quality mask from an input fringe pattern image. A novel loss function is proposed, large-scale (28 800 samples) real dataset collected to train the network. The experiments demonstrate proposed method achieves significantly improved accuracy overcomes main limitations of Fourier transform profilometry. In addition, excellent performance for...
Abstract A surge in technological advancements and innovations has spurred the rise of on‐demand meal delivery platforms. Despite their widespread appeal, these platforms face two critical challenges (i.e., order batching demand allocation) effectively managing process while maintaining reliability. In response, this study aims to address by examining effects real‐time demands restaurant density on reliability, as well how type driver in‐house versus crowdsourced drivers) moderates effects....
Knowledge Graphs (KGs) play a crucial role in enhancing e-commerce system performance by providing structured information about entities and their relationships, such as complementary or substitutable relations between products product types, which can be utilized recommender systems. However, relation labeling KGs remains challenging task due to the dynamic nature of domains associated cost human labor. Recently, breakthroughs Large Language Models (LLMs) have shown surprising results...
Graphene oxide nanosheets was prepared, functionlized with 3-aminopropyltriethoxysilane and used as the carrier for immobilization of papain. A high yield efficiency higher than 80% were achieved at pH 8.0 35 °C. The thermal stability immobilized papain improved compared free enzyme. And retained about 63% its initial activity after 30 days storage 4 °C, while enzyme only 41% same conditions. displayed binding affinity to substrate because unique properties GO. These results indicate that...
Recently, Graph Neural Networks (GNNs) have proven their effectiveness for recommender systems. Existing studies applied GNNs to capture collaborative relations in the data. However, real-world scenarios, a recommendation graph can be of various kinds. For example, two movies may associated either by same genre or director/actor. If we use single elaborate all these relations, too complex process. To address this issue, bring idea pre-training process step step. Based on divide-and-conquer,...