A5100373159- Logic, Reasoning, and Knowledge
- Topic Modeling
- Semantic Web and Ontologies
- Advanced Database Systems and Queries
- Advanced Text Analysis Techniques
- Multi-Agent Systems and Negotiation
- Logic, programming, and type systems
- Multimodal Machine Learning Applications
- Recommender Systems and Techniques
- Complex Network Analysis Techniques
- Advanced Optical Network Technologies
- Advanced Computational Techniques and Applications
- Natural Language Processing Techniques
- Advanced Graph Neural Networks
- Data Management and Algorithms
- Domain Adaptation and Few-Shot Learning
- Data Quality and Management
- Caching and Content Delivery
- Advanced Photonic Communication Systems
- Optical Network Technologies
- Advanced Malware Detection Techniques
- Sentiment Analysis and Opinion Mining
- Text and Document Classification Technologies
- Network Security and Intrusion Detection
- Cloud Data Security Solutions
Tsinghua University
2006-2024
Sichuan University
2024
National University of Defense Technology
2023
Nanjing University of Information Science and Technology
2022-2023
Zhejiang Lab
2023
Soochow University
2022-2023
Tianjin University
2017-2022
Beijing University of Technology
2019-2022
Nanjing University of Science and Technology
2019-2022
Dalian Jiaotong University
2022
Image emotion classification is an important computer vision task to extract emotions from images. The methods for image (IEC) are primarily based on label or distribution as a supervision signal, which neither has enough accessibility nor diversity, limiting the development of IEC research. Inspired by psychology research and recent booming large-scale pretrained language models. We figure out language-supervised paradigm, can cleverly combine features visual drive model gain stronger...
This paper analyses the connection availability of a protection paradigm, termed permanent dedicated path (P-DPP), in WDM mesh networks. An analytical model for calculating unavailability P-DPP is introduced. Numerical results on COST239 European network show that average proportional to blocking probability network, and fits well with simulation results.
Android access control granularity based on its permission mechanism is relatively coarse, which cannot effectively protect the user privacy. Many applications do not strictly abide by principle of least privilege (PLP). Both benign and malicious apps may request more permissions than those they really use. We rethink previous over-claim problem applications, extend it to three kinds problems: Explicit Permission Over-claim, Implicit Over-claim Ad Library Over-claim. The latter two problems...
Old photos are an important carrier to preserve the past. Usually, degradation of old is rather diverse and complex. Therefore, existing methods solve conventional restoration tasks difficult generalize. To this problem, we propose a novel method based on generative adversarial network. Our utilizes class-attributes complete in latent space. Specifically, divide process restoring into two stages, one global defect stage other local detail stage. In stage, extract representations four classes...
Abstract Recently, all kinds of geological disasters happen frequently on the earth. In China, there are countless earthquakes every year, which greatly affect country’s economic level and development as well people’s life health. The analysis seismic activity is becoming more significant. this article, spatial distribution China’s activities was analyzed by using provincial data from 1970 to 2013. On basis autocorrelation theory, Global Moran’s I , Local Indicators Spatial Association used...
As the usage of Cloud proliferates, need for security evaluation also grows. The process threat modeling and analysis is advocated to assess potential vulnerabilities that can undermine goals. However, given plethora distinct services involved in ecosystem varied attack surfaces entailed Cloud-specific architectures, performing a challenging task. Consequently, contemporary approaches, typically using relational models (e.g., graphs, trees...), primarily focus on specific services/layers...
Compared with standard text, understanding dialogue is more challenging for machines as the dynamic and unexpected semantic changes in each turn. To model such inconsistent semantics, we propose a simple but effective Hierarchical Dialogue Understanding model, HiDialog. Specifically, first insert multiple special tokens into turn-level attention to learn turn embeddings hierarchically. Then, heterogeneous graph module leveraged polish learned embeddings. We evaluate our on various tasks...
This paper applies a neural-network-based approximate dynamic programming (ADP) method, namely, the action dependent heuristic (ADHDP), to an industrial sucrose crystallization optimal control problem. The is nonlinear and slow time-varying process. It quite difficult establish precise mechanism model of because complex internal interacting variables. We developed neural network based on data from actual sugar boiling process factory. ADHDP learning- approximation-based approach which can...