- Service-Oriented Architecture and Web Services
- Topic Modeling
- Business Process Modeling and Analysis
- Natural Language Processing Techniques
- Mobile Agent-Based Network Management
- Semantic Web and Ontologies
- Advanced Graph Neural Networks
- Advanced Database Systems and Queries
- Petri Nets in System Modeling
- Peer-to-Peer Network Technologies
- Caching and Content Delivery
- Recommender Systems and Techniques
- Network Security and Intrusion Detection
- Software Engineering Research
- Software System Performance and Reliability
- Anomaly Detection Techniques and Applications
- Logic, Reasoning, and Knowledge
- Internet Traffic Analysis and Secure E-voting
- Web Data Mining and Analysis
- Multimedia Communication and Technology
- Collaboration in agile enterprises
- Advanced biosensing and bioanalysis techniques
- IPv6, Mobility, Handover, Networks, Security
- Simulation Techniques and Applications
- Technology and Security Systems
Peking University
2011-2025
Civil Aviation University of China
2024
Southwest University
2024
Beijing University of Posts and Telecommunications
2019-2020
Shandong University of Technology
2014
Sichuan Agricultural University
2012-2013
Institute of Software
2009-2011
Shaanxi Normal University
2007-2011
China University of Mining and Technology
2011
Hebei Normal University
2008
As the era of cloud technology arises, more and people are beginning to migrate their applications personal data cloud. This makes web-based an attractive target for cyber-attacks. a result, now need protections than ever. However, current anomaly-based web attack detection approaches face difficulties like unsatisfying accuracy lack generalization. And rule-based can hardly fight unknown attacks is relatively easy bypass. Therefore, we propose novel deep learning approach detect anomalous...
This study explores how to enhance the reasoning capabilities of large language models (LLMs) in knowledge base question answering (KBQA) by leveraging Monte Carlo Tree Search (MCTS). Semantic parsing-based KBQA methods are particularly challenging as these approaches require locating elements from bases and generating logical forms, demanding not only extensive annotated data but also strong capabilities. Although recent LLMs agents have demonstrated considerable potential, studies...
Low-frequency word prediction remains a challenge in modern neural machine translation (NMT) systems. Recent adaptive training methods promote the output of infrequent words by emphasizing their weights overall objectives. Despite improved recall low-frequency words, precision is unexpectedly hindered Inspired observation that form more compact embedding space, we tackle this from representation learning perspective. Specifically, propose frequency-aware token-level contrastive method, which...
News modeling and user are the two core tasks of news recommendation. Accurate representation can enable recommendation system to provide users with precise services. Most existing methods use deep learning models such as CNN Self-Attention extract text features from titles abstracts generate specific vectors. However, CNN-based have fixed parameters cannot for different input words, while Self-Attention-based high computational costs difficult capture local effectively. In our proposed...
News recommendation is of vital importance to alleviating in-formation overload. Recent research shows that precise modeling news content and user interests become critical for rec-ommendation. Existing methods usually utilize information such as title, abstract, entities predict Click Through Rate(CTR) or add some auxiliary tasks a multi-task learning framework. However, none them directly consider predicted popularity the degree users' attention popular into CTR prediction results....
In abstractive single-document summarization task, generated summaries always suffer from fabricated and less informative content. An intuitive way to alleviate this problem is merge external semantic knowledge into the model framework. paper, we incorporate explicit graphs based on knowledge, including term frequency, discourse information, entities with their relations, neural for problem. We propose a novel Summarization Semantic Knowledge Graphs (SKGSUM), which regards sentences as...
Abstract Natural language inference (NLI) is the basic task of many applications such as question answering and paraphrase recognition. Existing methods have solved key issue how NLI model can benefit from external knowledge. Inspired by this, we attempt to further explore following two problems: (1) make better use knowledge when total amount constant (2) bring more conveniently in application scenario. In this paper, propose a novel joint training framework that consists modified graph...
Social proverbs contain life philosophies and experience as well moral standards; aspects of social are reflected in the mirror proverbs.Social both language culture.Because their abundant cultural information, have been studied from point culture many researches.Modern semiotics has widely applied to studies an effective approach whose theories methods can be generally successfully used interpret meaning process various social-cultural phenomena.Bakhtinian semiotic theory seizes a crucial...
This paper proposes a practical static analysis tool named LUKE, for detecting null pointer dereferences (NPD) in C programs. LUKE first uses guarded value-dependence graph (VDG) to track the dependence relations of values, and then detects NPD by solving reachability problem on its VDG. To improve accuracy as well scalability, detection algorithm leverages heuristic inference algorithms results control dependences analysis. We evaluated 10 large-scale open source projects, show that has...
An intruder of a company’s network may use stolen login credentials to silently collect sensitive data. Such malicious user behavior is difficult detect as long it does not trigger access violation or data leak alert. In this paper, we propose an ensemble three unsupervised anomaly detection algorithms, namely OCSVM, RNN and Isolation Forest, abnormal patterns. Besides, User Behavior Analytics (UBA) Platform proposed logs, extract features conduct experiments. The experiment results indicate...
Social proverbs contain life philosophies and experience as well moral standards; aspects of social are reflected in the mirror proverbs.Social both language culture.Because their abundant cultural information, have been studied from point culture many researches.Although researches done to reveal similarities differences between English Chinese proverbs, they hardly focus on deep structure or refracted them.A comparative study by illustrating connection with structures, is provided.The...
Passive RFID event refers to one kind of composite in middleware, or more constituent sub-events which do not occur under certain condition. Definition passive requires careful consideration temporal constraint and composition hierarchy, moreover, the detection well-organized query plan for eventpsilas absence runtime context. To process complex business logic, we should be able define a wide variety types with rigorous semantics as well perform efficient detection. In this paper, first...
Traditional application integration technologies are performed in a rigid and slow process do not fully utilize the computing power storage capability of client. By extending mashup concept into space, we can achieve novel more lightweight approach. This paper introduces our Web based platform BU Studio, which is built upon series abstractions. These abstractions proposed perspective software architecture, including component model to encapsulate integrated objects connector specify...
For performance and security concern in large-scale application environment, based on "follow the chain" mode, a distributed RFID discovery system is proposed implemented. At first, architecture given. Secondly, core components, such as query engine, local cache client, are formally specified with TIOA. Experiments execution time accuracy show that this more suitable for applications. The also verified by practical project "Beijing Volunteer Card Management Platform".
The coal mining mode of paste-like fill shows the connotation and characteristics green both in filling materials backfill technology. Combined with practical situation colliery, some main components are introduced emphatically, such as cemented material, preparation transportation materials, technology at working face; application prospects on analyzed.
In order to build large-scale RFID applications, it is necessary establish public service infrastructure, the essential task efficient and secure code resolution network. Aiming at alleviating problems in load balancing, single node failure tolerance security of network, based on related standards research works, a hierarchical P2P network structure proposed, k-Nearest Neighbor storing querying mechanism are presented static resolution, secondary indexing dynamic double random number entity...
Executable modeling allows the models to be executed and treated as prototypes determine behaviors of a system. In this paper, we propose an approach for formalizing requirement generating executable from them. Application activity diagrams (AADs), which are used represent dynamic systems in capability requirements models, firstly formalized saved XML documents. Then, on basis these mapping algorithm translating AADs into instances simulation is proposed. A case study finally given...
In order to build open-loop large scale RFID applications, it is necessary establish public service infrastructure, the key issue efficient and robust code resolution network. But there exist some problems, such as load balancing, single node failure, etc. On basis of recent research, according features governmental administration enterprise a hierarchical P2P based RFDD network structure proposed implemented. The core components are specified formally by TIOA. Meanwhile, management toolset...