A5028258863- Complex Network Analysis Techniques
- Advanced Graph Neural Networks
- Topic Modeling
- Natural Language Processing Techniques
- Graph Theory and Algorithms
- Topological and Geometric Data Analysis
- Semantic Web and Ontologies
- Advanced Steganography and Watermarking Techniques
- Bayesian Modeling and Causal Inference
- Opinion Dynamics and Social Influence
- Chaos-based Image/Signal Encryption
- Speech and dialogue systems
- Digital Media Forensic Detection
- Advanced Database Systems and Queries
- Biomedical Text Mining and Ontologies
- Network Security and Intrusion Detection
- Homotopy and Cohomology in Algebraic Topology
- Algorithms and Data Compression
- Advanced Algebra and Geometry
- Neural Networks and Applications
- Age of Information Optimization
- Collaboration in agile enterprises
- Land Use and Ecosystem Services
- Image Retrieval and Classification Techniques
- Domain Adaptation and Few-Shot Learning
China University of Mining and Technology
2008-2025
Nanyang Technological University
2016-2025
Chongqing Normal University
2024
Zhejiang University
2003-2021
Alibaba Group (United States)
2020-2021
Fudan University
2010-2020
Massachusetts Institute of Technology
2020
Chinese Academy of Fishery Sciences
2012-2018
Tokyo University of Information Sciences
2018
The University of Tokyo
2018
We study the problem of identifying infection sources in a network based on topology, and subset timestamps. In case single source tree network, we derive maximum likelihood estimator unknown diffusion parameters. then introduce new heuristic involving an optimization over parametrized family Gromov matrices to develop estimation algorithm for general graphs. Compared with breadth-first search commonly adopted literature, simulations demonstrate that our approach achieves better accuracy...
We study the problem of identifying multiple rumor or infection sources in a network under susceptibleinfected model, and where these may start spreading at different times.We introduce notion an abstract estimator, which given graph, assigns higher value to each vertex graph it considers more likely be source.This includes several single-source estimators developed literature.We concepts quasi-regular tree heavy center, allows us develop algorithmic framework that transforms estimator into...
In the context of digital transformation, varying dimensions maturity significantly influence value creation enhancement for enterprises. Optimizing these to augment corporate represents an urgent challenge manufacturing This study examines 355 listed automotive enterprises (including auto parts and related businesses) through multi-case analysis, grounded theory, QCA methodology investigate intrinsic mechanisms pathways linking transformation with in manufacturing. The sample were...
Graph signal processing (GSP) has become an important tool in many areas such as image processing, networking learning and analysis of social network data. In this paper, we propose a broader framework that not only encompasses traditional GSP special case, but also includes hybrid graph classical over continuous domain. Our relies extensively on concepts tools from functional to generalize signals separable Hilbert space with infinite dimensions. We develop concept analogous Fourier...
Graph signal processing (GSP) uses a shift operator to define Fourier basis for the set of graph signals. The is often chosen capture topology. However, in many applications, topology may be unknown <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a priori</i> , its structure uncertain, or generated randomly from predefined each observation. Each gives rise different operator. In this paper, we develop GSP framework over probability space...
The theory of sampling and recovery bandlimited graph signals has been extensively studied. However, in many cases, the observation a signal is quite coarse. For example, users only provide simple comments such as “like” or “dislike” for product on an e-commerce platform. This particular scenario where sign information can be measured. In this paper, we are interested how to sample based online manner, by which direction original estimated. signed problem formulated Markov decision process...
This paper studies learning from positive and unlabeled examples, known as PU learning. It proposes a novel method called Predictive Adversarial Networks (PAN) based on GAN (Generative Networks). learns generator to generate data (e.g., images) fool discriminator which tries determine whether the generated belong (positive) training class. can be casted trying identify (not generate) likely instances set that determines identified are indeed positive. However, directly applying is...
Robust lossless data hiding (LDH) methods have attracted more and attentions for copyright protection of multimedia in lossy environment. One the important requirements robust LDH is reversibility, that is, host images can be recovered without any distortion after hidden messages are removed. The reversibility often guaranteed by embedding model, which affects performance greatly. Another requirement to a better robustness, allows well adaptable environment, e.g., JPEG compression. In this...
The spanning tree heuristic is a commonly adopted procedure in network inference and estimation. It allows one to generalize an method developed for trees, which usually based on statistically rigorous approach, general graphs by (usually randomly) choosing the graph apply approach trees. However, there are intractable number of trees dense graph. In this paper, we represent weighted with matrix, call Gromov matrix. We propose that constructs family matrices using convex combinations, can be...
In our digital era, platforms become an unprecedented opportunities for manufacturing enterprises to leverage their business strategy. This study investigates how can enhance organizational duality through platforms. Specifically, this examines the effect of and capability reconfiguration on entrepreneurial duality. The also network atmosphere moderates relationship. Based analysis 286 undergoing transformation, results indicate that positively affect both reconfiguration. have a positive...
In clustering algorithms, the selection of neighbors significantly affects quality final results. While various neighbor relationships exist, such as <i>K-</i>nearest neighbors, natural and shared most can only handle single structural relationships, identification accuracy is low for datasets with multiple structures. life, people's first instinct complex things to divide them into parts complete. Partitioning dataset more sub-graphs a good idea approach identifying Taking inspiration from...
This paper proposes a hybrid basis function construction method (GP-RVM) for Symbolic Regression problem, which combines an extended version of Genetic Programming called Kaizen and Relevance Vector Machine to evolve optimal set functions. Different from traditional evolutionary algorithms where single individual is complete solution, our solution based on linear combination functions built individuals during the evolving process. RVM sparse Bayesian kernel selects suitable constitute basis....
This paper proposes a hybrid basis function construction method (GP-RVM) for Symbolic Regression problem, which combines an extended version of Genetic Programming called Kaizen and Relevance Vector Machine to evolve optimal set functions. Different from traditional evolutionary algorithms where single individual is complete solution, our solution based on linear combination functions built individuals during the evolving process. RVM sparse Bayesian kernel selects suitable constitute basis....
Wind power point forecasting is the primary method to deal with its uncertainty.However, in many applications, probabilistic interval of wind more useful than traditional forecasting.Methods determine value very essential system operations.Based on bootstrap method, this paper proposed a wavelet transform combined neuro-fuzzy network model estimate prediction power.In model, account for ramp event series, wavelet-based was used and moving block which considers dependence construct sampling...