A5100451429- Data Management and Algorithms
- Complex Network Analysis Techniques
- Graph Theory and Algorithms
- Recommender Systems and Techniques
- Nonlinear Photonic Systems
- Nonlinear Waves and Solitons
- Advanced Graph Neural Networks
- Network Security and Intrusion Detection
- Spam and Phishing Detection
- Advanced Database Systems and Queries
- Internet Traffic Analysis and Secure E-voting
- Advanced Image and Video Retrieval Techniques
- Online Learning and Analytics
- Automated Road and Building Extraction
- Advanced Mathematical Physics Problems
- Metal Extraction and Bioleaching
- Caching and Content Delivery
- Extraction and Separation Processes
- Advanced Bandit Algorithms Research
- Face and Expression Recognition
- Innovation and Knowledge Management
- Fractional Differential Equations Solutions
- Transportation Planning and Optimization
- Topological and Geometric Data Analysis
- Blind Source Separation Techniques
University of Hong Kong
2023-2025
Wuhan Botanical Garden
2025
Shenzhen University
2025
Chinese Academy of Sciences
2025
University of Leicester
2025
Hong Kong University of Science and Technology
2023-2025
Ocean University of China
2024
China State Shipbuilding (China)
2024
Ningbo University
2019-2024
Beijing National Laboratory for Molecular Sciences
2024
Weanling pigs, with immature immune system and physiological function, usually experience post-weaning diarrhea. This study determined the effects of dietary Clostridium butyricum supplementation on growth performance, diarrhea, immunity weaned pigs challenged lipopolysaccharide (LPS).In Experiment (Exp.) 1, 144 piglets were at 21 d randomly assigned to six groups, replicates per group four replicate, receiving a control diet (CON) or supplemented antibiotics (AB) C. (CB) (0.1%, 0.2%, 0.4%,...
Mobile Edge Computing (MEC) is a promising architecture to reduce the energy consumption of mobile devices and provide satisfactory quality-of-service time-sensitive services. How jointly optimize task offloading resource allocation minimize subject latency requirement remains an open problem, which motivates this paper. When constraint taken into account, optimization variables, including ratio, transmission power, subcarrier computing allocation, are strongly coupled. To address issue, we...
While massive open online courses (MOOCs) have gained increasing popularity in recent years, dropout prediction has been an important task to solve due the high rates of students found MOOCs. Current methods normally apply supervised learning prediction, using general features extracted from behavior records without concern types. However, student is diverse and there are no sufficient labeled data train a model because it time-costing label enormous practice. To these problems, this work,...
Based on the hybrid solutions to (2+1)-dimensional Kadomtsev–Petviashvili (KP) equation, motion trajectory of KP equation is further studied. We obtain a single lump before and after collision with line, lump, breather waves by approximating along some parallel orbits at infinity. derive mathematical expression phase change wave. At same time, we give plots reveal obvious change. Our method proposed find wave can be applied other integrable equations. The results expand understanding...
Distance labeling approaches are widely adopted to speed up the online performance of shortest distance queries. The construction labeling, however, can be exhaustive especially on big graphs. For a major category large graphs, small-world networks, state-of-the-art approach is Pruned Landmark Labeling (PLL). PLL prunes labels based node order and directly constructs pruned by performing breadth-first searches in order. pruning technique, as well index construction, has strong sequential...
Hypergraphs provide a versatile framework for modeling complex relationships beyond pairwise interactions, finding applications in various domains. k -core decomposition is fundamental task hypergraph analysis that decomposes hypergraphs into cohesive substructures. Existing studies capture the cohesion based on vertex neighborhood size. However, such poses unique challenges, including efficiency of core value updates, redundant computation, and high memory consumption. We observe...
Collaborative filtering (CF) has been widely used in recommender systems to generate personalized recommendations. However, using CF are vulnerable shilling attacks, which attackers inject fake profiles manipulate recommendation results. Thus, attacks pose a threat the credibility of systems. Previous studies mainly derive features from characteristics item ratings user detect attackers, but methods suffer low accuracy when adopt new rating patterns. To overcome this drawback, we properties...
Computing the shortest path between two vertices is a fundamental problem in road networks. Most of existing works assume that edges networks have no labels, but many real applications, labels and label constraints may be placed on appearing valid path. Hence, we study label-constrained queries this paper. In order to process such efficiently, adopt an index-based approach propose novel index structure, LSD-Index, based tree decomposition. With design efficient query processing algorithm...
The acquisition of explicit user feedback (e.g., ratings) in real-world recommender systems is often hindered by the need for active involvement. To mitigate this issue, implicit clicks) generated during browsing exploited as a viable substitute. However, possesses high degree noise, which significantly undermines recommendation quality. While many methods have been proposed to address issue assigning varying weights feedback, two shortcomings persist: (1) weight calculation these...
Biomimetic soft robots are typically made of materials with bioinspired configurations. However, their locomotion is activated and manipulated by externally controlled actuators. In this study, piezo-wormbots were developed automatically triggering the mechanical metamaterial-inspired actuator to mimic continuous crawling inchworms without manipulation, where was deformation themselves. We designed flexible an generate bending under inflation, piezoelectric rubber create internal excitation...
Acquiring and representing the 4D space of rays in world (the light field) is important for many computer vision graphics applications. In this paper, we propose an iterative method to acquire field from a focal stack. First, discrete refocusing equation derived integral imaging principles. With equation, linear projection system formulated model stack process. Then reconstruct through solving inverse problem with filtering-based method. The experimental results show that our approach...
In indexing a graph for distance queries, labeling is common practice; in particular, 2-hop which guarantees the exactness of query results widely adopted. When it comes to massive real with relatively large treewidth such as social networks and web graphs, however, can hardly be constructed due oversized index. This paper discloses theoretical relationships between labeling's index size time. To scale up labeling, this proposes Core-Tree (CT) Index facilitate critical effective trade-off...
The shortest-path query, which returns the shortest path between two vertices, is a basic operation on complex networks and has numerous applications. To handle queries, one option to use traversal-based methods (e.g., breadth-first search); another extension-based methods, i.e., extending existing that indexes shortest-distance queries support queries. These types of make different trade-offs in query time space cost, but comprehensive studies their performance real-world graphs are...
Fine-grained object detection (FGOD) extends with the capability of fine-grained recognition. In recent two-stage FGOD methods, region proposal serves as a crucial link between and However, current methods overlook that some proposal-related procedures inherited from general are not equally suitable for FGOD, limiting multitask learning generation, representation, to utilization. this article, we present enhancement (PETDet) better handle subtasks in methods. First, an anchor-free...
Traditional recommender systems assume that all the users are independent, and they usually face cold start data sparse problems. To alleviate these problems, social use relations as an additional input to improve recommendation accuracy. Social follows intuition people with relationships share some kinds of preference towards items. Current methods commonly apply Matrix Factorization (MF) model incorporate information into process. As alternative MF, we propose a novel approach based on...
This paper presents a method for the simultaneous determination of 48 fragrance allergens in four types toys (plastic toys, play clays, plush and toys) based on GC with ion trap MS/MS. Compared single-stage MS, MS/MS is superior terms qualification quantification large range compounds complicated matrices. Procedures extraction purification were optimized each toy type. The proved to be linear over wide concentrations all analytes correlation coefficients between 0.9768 0.9999. Validation...
In this paper, the (2+1)-dimensional fifth-order KdV equation is analytically investigated. By using Hirota’s bilinear method combined with perturbation expansion, high-order breather solutions of are generated. Then, lump also derived from soliton by a long-wave limit and some suitable parameter constraints. Furthermore, we extend to obtain hybrid taking for partial solutions. Finally, dynamic behavior these presented in figures.