5分で分かる!? 有名論文ナナメ読み：Jacob Devlin et al. : BERT : Pre-training of Deep Bidirectional Transformers for Language Understanding

The storage of electric energy is of ever growing importance for our modern, technology-based society, and novel battery systems are in the focus of research. The substitution of conventional metals as redox-active material by organic materials offers a promising alternative for the next generation of rechargeable batteries since these organic batteries are excelling in charging speed and cycling stability. This review provides a comprehensive overview of these systems and discusses the numerous classes of organic, polymer-based active materials as well as auxiliary components of the battery, like additives or electrolytes. Moreover, a definition of important cell characteristics and an introduction to selected characterization techniques is provided, completed by the discussion of potential socio-economic impacts.

Polymer-Based Organic Batteries

Malicious URL, a.k.a. malicious website, is a common and serious threat to cybersecurity. Malicious URLs host unsolicited content (spam, phishing, drive-by exploits, etc.) and lure unsuspecting users to become victims of scams (monetary loss, theft of private information, and malware installation), and cause losses of billions of dollars every year. It is imperative to detect and act on such threats in a timely manner. Traditionally, this detection is done mostly through the usage of blacklists. However, blacklists cannot be exhaustive, and lack the ability to detect newly generated malicious URLs. To improve the generality of malicious URL detectors, machine learning techniques have been explored with increasing attention in recent years. This article aims to provide a comprehensive survey and a structural understanding of Malicious URL Detection techniques using machine learning. We present the formal formulation of Malicious URL Detection as a machine learning task, and categorize and review the contributions of literature studies that addresses different dimensions of this problem (feature representation, algorithm design, etc.). Further, this article provides a timely and comprehensive survey for a range of different audiences, not only for machine learning researchers and engineers in academia, but also for professionals and practitioners in cybersecurity industry, to help them understand the state of the art and facilitate their own research and practical applications. We also discuss practical issues in system design, open research challenges, and point out some important directions for future research.

Malicious URL Detection using Machine Learning: A Survey

The European Telecommunications Standards Institute (ETSI) has introduced the paradigm of Multi-Access Edge Computing (MEC) to enable efficient and fast data processing in mobile networks. Among other technological requirements, security and privacy are significant factors in the realization of MEC deployments. In this paper, we analyse the security and privacy of the MEC system. We introduce a thorough investigation of the identification and the analysis of threat vectors in the ETSI standardized MEC architecture. Furthermore, we analyse the vulnerabilities leading to the identified threat vectors and propose potential security solutions to overcome these vulnerabilities. The privacy issues of MEC are also highlighted, and clear objectives for preserving privacy are defined. Finally, we present future directives to enhance the security and privacy of MEC services.

/pdf/survey-on-multi-access-edge-computing-security-and-privacy-5epo4z3o46.pdf

Survey on Multi-Access Edge Computing Security and Privacy

We introduce a new method for triggering vulnerabilities in deep layers of binary executables and facilitate their exploitation. In our approach we combine dynamic symbolic execution with fuzzing techniques. To maximize both the execution path depth and the degree of freedom in input parameters for exploitation, we define a novel method to assign probabilities to program paths. Based on this probability distribution we apply new path exploration strategies. This facilitates payload generation and therefore vulnerability exploitation.

Detection of Intrusions and Malware, and Vulnerability Assessment

Anomaly detection generally involves the extraction of features from entities' or users' properties, and the design of anomaly detection models using machine learning or deep learning algorithms. However, only considering entities' property information could lead to high false positives. We posit the importance of also considering connections or relationships between entities in the detecting of anomalous behaviors and associated threat groups. Therefore, in this paper, we design a GCN (graph convolutional networks) based anomaly detection model to detect anomalous behaviors of users and malicious threat groups. The GCN model could characterize entities' properties and structural information between them into graphs. This allows the GCN based anomaly detection model to detect both anomalous behaviors of individuals and associated anomalous groups. We then evaluate the proposed model using a real-world insider threat data set. The results show that the proposed model outperforms several state-of-art baseline methods (i.e., random forest, logistic regression, SVM, and CNN). Moreover, the proposed model can also be applied to other anomaly detection applications.

Anomaly Detection with Graph Convolutional Networks for Insider Threat and Fraud Detection

Online social media promotes the development of the news industry and make it easy for everyone to obtain the latest news. Meanwhile, the circumstances get worse because of fake news. Fake news is flooding and become a serious threat which may cause high societal and economic losses, making fake news detection important. Unlike traditional one, news on social media tends to be short and misleading, which is more confusing to identify. On the other hand, fake news may contain parts of the facts and parts of the incorrect contents in one statement, which is not so clear and simple to classify. Hence, we propose a two-stage model to deal with the difficulties. Our model is built on BERT, a pre-trained model with a more powerful feature extractor Transformer instead of CNN or RNN. Besides, some accessible information is used to extend features and calculate attention weights. At last, inspired by fine-grained sentiment analysis, we treat fake news detection as fine-grained multiple-classification task and use two similar sub-models to identify different granularity labels separately. We evaluate our model on a real-world benchmark dataset. The experimental results demonstrate its effectiveness in fine-grained fake news detection and its superior performance to the baselines and other competitive approaches.

A two-stage model based on BERT for short fake news detection

URL and DNS are two common attack vectors in malicious network activities; thus, detection for malicious URL and DNS is crucial in network security. In this paper, we propose an online detection scheme based on character-level deep neural networks. Specifically, this scheme maps the URL and DNS strings into vector form using some natural language processing methods. The CNN (Convolutional Neural Network) network framework is then designed to automatically extract the malicious features and train the classifying model. Experimental results on real-world URL and DNS datasets show that proposed method outperforms several state-of-art baseline methods, in terms of efficiency and scalability.

A Deep Learning Based Online Malicious URL and DNS Detection Scheme

Distributed Denial of Service (DDoS) has been one of the most critical threats to internet applications and web services Especially with the current advances in network technology, many attackers resort to application layer DDoS (ALDDoS) which utilizes legitimate requests to overwhelm the victim servers Under this kind of attack, the single request content can be highly similar to normal ones, and this renders previous traffic features-based detection methods void In this paper, we are addressing two common issues in ALDDoS detection methods: the inaccuracy of traffic feature based detecting algorithms, and the time and space complexity of user behavior-based detecting algorithms Different from the existing detection pattern for each request, the detection pattern used in this paper is for a time window We extract instances of traffic and user behaviors from web server logs, and propose a hybrid traffic-user behavior detection method for ALDDoS Neutral network is adopted for further cluster analysis Experimental results on the recent public dataset CICIDS2017 indicate that the proposed method can achieve high detection accuracy while reducing 90% of time cost

ALDD: A Hybrid Traffic-User Behavior Detection Method for Application Layer DDoS

Due to rich characteristics and functionalities, PDF format has become the de facto standard for the electronic document exchange. As vulnerabilities in the major PDF viewers have been disclosed, a number of methods have been proposed to tame the increasing PDF threats. However, one recent evasion exploit is found to evade most of detections and renders all of the major static methods void. Moreover, many existing vulnerabilities identified before can now evade the detection through exploiting this evasion exploit. In this paper, we introduce this newly identified evasion exploit and propose a new feature extractor FEPDF to detect malicious PDFs. Based on the FEPDF and the JavaScript detection model, we test the performance of the proposed feature extractor FEPDF, and evaluation results show that FEPDF has a satisfactory performance in malicious PDF detection.

Chao Liu

Papers

Anomaly Detection with Graph Convolutional Networks for Insider Threat and Fraud Detection

A two-stage model based on BERT for short fake news detection

A Deep Learning Based Online Malicious URL and DNS Detection Scheme

ALDD: A Hybrid Traffic-User Behavior Detection Method for Application Layer DDoS

FEPDF: A Robust Feature Extractor for Malicious PDF Detection