A5100350503- Complex Network Analysis Techniques
- Caching and Content Delivery
- Peer-to-Peer Network Technologies
- Human Mobility and Location-Based Analysis
- Spam and Phishing Detection
- Opinion Dynamics and Social Influence
- Topic Modeling
- Network Security and Intrusion Detection
- Internet Traffic Analysis and Secure E-voting
- Advanced Graph Neural Networks
- Data Visualization and Analytics
- Multimodal Machine Learning Applications
- Opportunistic and Delay-Tolerant Networks
- Network Traffic and Congestion Control
- Advanced Malware Detection Techniques
- Natural Language Processing Techniques
- IoT and Edge/Fog Computing
- Digital Marketing and Social Media
- Domain Adaptation and Few-Shot Learning
- Anomaly Detection Techniques and Applications
- Advanced Text Analysis Techniques
- Misinformation and Its Impacts
- AI in cancer detection
- Advanced Neural Network Applications
- Software-Defined Networks and 5G
Tongji University
2017-2025
Fudan University
2016-2025
Simcere Pharmaceutical (China)
2023-2025
Xidian University
2007-2025
Chongqing University of Posts and Telecommunications
2025
Liaoning University of Traditional Chinese Medicine
2025
Shanghai University
2021-2025
Sichuan Agricultural University
2025
Zhejiang University of Science and Technology
2025
University of Florida
2025
Language is essentially a complex, intricate system of human expressions governed by grammatical rules. It poses significant challenge to develop capable AI algorithms for comprehending and grasping language. As major approach, language modeling has been widely studied understanding generation in the past two decades, evolving from statistical models neural models. Recently, pre-trained (PLMs) have proposed pre-training Transformer over large-scale corpora, showing strong capabilities...
Background: Treatment of cancers with programmed cell death protein 1 (PD-1) pathway inhibitors can lead to immune-related adverse events (irAEs), which could be serious and even fetal. Therefore, clinicians should aware the characteristics irAEs associated use such drugs. Methods: The MEDLINE, EMBASE, Cochrane databases were searched find potential studies using following strategies: anti-PD-1/PD-L1 treatment; irAEs; cancer. R© package Meta was used pool incidence. Results: Forty-six...
The measurement scales, consistency index, inconsistency issues, missing judgment estimation and priority derivation methods have been extensively studied in the pairwise comparison matrix (PCM). Various approaches proposed to handle these problems, made great contributions decision making. This paper reviews literature of main developments PCM. There are plenty related thus we mainly focus on published 37 peer reviewed international journals from 2010 2015 (searched via ISI Web science). We...
The evaluation of feature selection methods for text classification with small sample datasets must consider performance, stability, and efficiency. It is, thus, a multiple criteria decision-making (MCDM) problem. Yet there has been few research in using MCDM which considering criteria. Therefore, we use MCDM-based evaluating datasets. An experimental study is designed to compare five validate the proposed approach 10 methods, nine measures binary classification, seven multi-class three...
Currently, online social networks such as Facebook, Twitter, Google+, LinkedIn, and Foursquare have become extremely popular all over the world play a significant role in people¿s daily lives. People access OSNs using both traditional desktop PCs new emerging mobile devices. With more than one billion users worldwide, are venue of innovation with many challenging research problems. In this survey, we aim to give comprehensive review state-of-the-art related user behavior from several...
Multiple instance learning (MIL) is a powerful tool to solve the weakly supervised classification in whole slide image (WSI) based pathology diagnosis. However, current MIL methods are usually on independent and identical distribution hypothesis, thus neglect correlation among different instances. To address this problem, we proposed new framework, called correlated MIL, provided proof for convergence. Based devised Transformer (TransMIL), which explored both morphological spatial...
In this paper, from the perspective of opinion dynamics theory, we investigate interaction mechanism a group autonomous agents in an e-commerce community (or social network), and influence power leaders during formation opinion. According to opinion's update manner influence, paper divides within network into two subgroups: followers. Then, establish new bounded confidence-based dynamic model for followers simulate evolution agents. Through numerical simulations, further opinion,...
As of June 1, 2020, coronavirus disease 2019 (COVID-19) has caused more than 6,000,000 infected persons and 360,000 deaths globally. Previous studies revealed pregnant women with COVID-19 had similar clinical manifestations to nonpregnant women. However, little is known about the outcome neonates born
Large language models (LLMs) such as ChatGPT have garnered global attention recently, with a promise to disrupt and revolutionize business operations. As managers rely more on artificial intelligence (AI) technology, there is an urgent need understand whether are systematic biases in AI decision-making since they trained human data feedback, both may be highly biased. This paper tests broad range of behavioral commonly found humans that especially relevant operations management. We although...
Being able to keep the graph scale small while capturing properties of original social graph, sampling provides an efficient, yet inexpensive solution for network analysis. The challenge is how create a small, but representative sample out massive with millions or even billions nodes. Several algorithms have been proposed in previous studies, there lacks fair evaluation and comparison among them. In this paper, we analyze state-of art evaluate their performance on some widely recognized...
Microblogging service has emerged to be a dominant web medium for billions of individuals sharing and spreading instant news information, therefore monitoring the event evolution on microblog sphere is crucial providing both better user experience deeper understanding real-time events. In this paper we explore problem generating storylines from microblogs input queries. This challenging due sparse, dynamic social nature microblogs. Given query an ongoing event, propose sketch storyline by...
Our daily lives have been immersed in widespread location-based social networks (LBSNs). As an open platform, LBSNs typically allow all kinds of users to register accounts. Malicious attackers can easily join and post misleading information, often with the intention influencing users' decisions urban computing environments. To provide reliable information improve experience for legitimate users, we design implement DeepScan, a malicious account detection system LBSNs. Different from existing...
Network coordinate (NC) systems provide a lightweight and scalable way for predicting the distances, i.e., round-trip latencies among Internet hosts. Most existing NC embed hosts into low dimensional Euclidean space. Unfortunately, persistent occurrence of Triangle Inequality Violation (TIV) on largely limits distance prediction accuracy those systems. Some alternative aim at handling TIV, however, they only achieve comparable with based In this paper, we propose an system, so-called...
In a Website Fingerprinting (WF) attack, local, passive eavesdropper utilizes network flow information to identify which web pages user is browsing. Previous researchers have extensively demonstrated the feasibility and effectiveness of WF, but only under strong Single Page Assumption: extracted by adversary always belongs single page. other words, WF classifier will never be asked classify corresponding more than one page, or part The Assumption unrealistic because people often browse with...
We describe Neutron, a version of the TinyOS operating system that efficiently recovers from memory safety bugs. Where existing schemes reboot an entire node on error, Neutron's compiler and runtime extensions divide programs into recovery units only faulting unit. The kernel itself is unit: violation appears to applications as processor being unavailable for 10-20 milliseconds.