Zhen Xu

Chinese Academy of Sciences | CAS · Institute of Information Engineering

Robust keypoint detection on omnidirectional images against large perspective variations, is a key problem in many computer vision tasks. In this paper, we propose a perspectively equivariant keypoint learning framework named OmniKL for addressing this problem. Specifically, the framework is composed of a perspective module and a spherical module,...

In a permissioned blockchain, performance dictates its development, which is substantially influenced by its parameters. However, research on auto-tuning for better performance has somewhat stagnated because of the difficulty posed by distributed parameters; thus, it is possible only with difficulty to propose an effective auto-tuning optimization...

The emergence of vertical federated learning (VFL) solves the problem of joint modeling between participants sharing the same ID space and different feature spaces. Privacy-preserving (PP) VFL is challenging because complete sets of labels and features are not owned by the same entity, and more frequent and direct interactions are required between...

Hunting the advanced threats hidden in the enterprise networks has always been a complex and difficult task. Due to the variety of attacking means, it is difficult for traditional security systems to detect threats. Most existing methods analyze log records, but the amount of log records generated every day is very large. How to find the informatio...

The application of deep neural networks (DNNs) on 360-degree images has achieved remarkable progress in the recent years. However, DNNs have been demonstrated to be vulnerable to well-crafted adversarial examples, which may trigger severe safety problems in the real-world applications based on 360-degree images. In this paper, we propose an adversa...

As essential work in IT operations, anomaly lo-calization, aiming to identify the affected scope of Internet infrastructure once an anomaly alarm occurs, is challenging due to the huge search space. The existing solutions usually show limited performances in the CDN scenario since they take the desirable assumptions that do not match with the pract...

The safety of Deep Neural Networks (DNNs) processing om-nidirectional images (ODIs) is an under-researched topic. In this paper, we propose a novel sparse attack, named Single-Perspective (SP) Attack, towards fooling these models by perturbing only one perspective image (PI) rendered from the target ODI. The attack is launched from the perspective...

The application of deep neural networks (DNNs) on 360-degree images has achieved remarkable progress in the recent years. However, DNNs have been demonstrated to be vulnerable to well-crafted adversarial examples, which may trigger severe safety problems in the real-world applications based on 360-degree images. In this paper, we propose an adversa...

With the rapid advancements in Augmented Reality, the number of AR users is gradually increasing and the multiuser AR ecosystem is on the rise. Currently, AR applications usually present results without limitations, which causes great latent danger to users, so it is necessary to apply strategies to ensure the safe output of AR. Due to the environm...

Long-term prediction is still a difficult problem in data mining. People usually use various kinds of methods of Recurrent Neural Network to predict. However, with the increase of the prediction step, the accuracy of prediction decreases rapidly. In order to improve the accuracy of long-term prediction,we propose a framework Variational Auto-Encode...

Rotational-XOR cryptanalysis is a very recent technique for ARX ciphers. In this paper, the probability propagation formula of RX-cryptanalysis in modular addition is extended, and the calculation of RX-difference probability for any rotation parameter ($0<k<n$) can be realized. By proposing a concept of RX-offset and constructing the corresponding...

Fake news with textual and visual contents has a better story-telling ability than text-only contents , and can be spread quickly with social media. People can be easily deceived by such fake news, and traditional expert identification is labor-intensive. Therefore, automatic detection of multimodal fake news has become a new hot-spot issue. A shor...

Mobile edge computing (MEC), extending computing services from cloud to edge, is recognized as one of key pillars to facilitate real-time services and tackle backhaul bottleneck. However, it is not economically efficient to attach intensive security appliances to every MEC node to defend application-level DDoS attacks and ensure the availability of...

With the capability of trading accuracy for latency on-the-fly, the technique of adaptive early-exit inference has emerged as a promising line of research to accelerate the deep learning inference. However, studies in this line of research commonly use a group of thresholds to control the accuracy-latency trade-off, where a thorough and general met...

Similar to conventional video, the increasingly popular 360 virtual reality (VR) video requires copyright protection mechanisms. The classic approach for copyright protection is the introduction of a digital watermark into the video sequence. Due to the nature of spherical panorama, traditional watermarking schemes that are dedicated to planar medi...

Multivariate Time series data play important roles in our daily life. How to use these data in the process of prediction is a highly attractive study for many researchers. To achieve this goal, in this paper, we present a novel multivariate time series prediction method based on multi-attention generative adversarial network. This method includes t...

Cyber-Physical-Social Systems are frequently prescribed for providing valuable information on personalized services. The foundation of these services is big data which must be trustily collected and efficiently processed. Though High Performance Computing and Communication technique makes great contributions to addressing the issue of data processi...

Delivering high fidelity virtual reality (VR) video over mobile networks is very challenging since VR applications usually require very high bandwidth and ultra low latency. With the evolution of 5G mobile networks, multi-cell multi-access edge computing (MEC) networks enable low latency data communication. However, even in this setting, the requir...

A primary concern in creating compartments (i.e., protection domains) for bare-metal systems is to adopt the applicable compartmentalization policy. Existing studies have proposed several typical policies in literature. However, neither of the policies consider the influence of unsafe functions on the compartment security that a vulnerable function...

Migration-based Dynamic Platform (MDP) technique, a type of Moving Target Defense (MTD) techniques, defends against sophisticated cyber-attacks by randomly and dynamically selecting a platform for executing service/job. Security defense mechanisms protect service/job usually at the cost of degrading its performance. Therefore, it is valuable to mak...

Inference based on deep learning models is usually implemented by exposing sensitive user data to the outside models, which of course gives rise to acute privacy concerns. To deal with these concerns, Dong et al. recently proposed an approach, namely the dropping-activation-outputs (DAO) first layer. This approach was claimed to be a non-invertib...

Automated malware classification using deep learning techniques has been widely researched in recent years. However, existing studies addressing this problem are always based on the assumption of closed world, where all the categories are known and fixed. Thus, they lack robustness and do not have the ability to recognize novel malware instances. I...

Software-Defined Networking (SDN) enables network operators to easily and quickly manage, configure, and optimize network resources via decoupling the control logic from forwarding devices. Such a separation, however, brings multiple potential bottlenecks that can be exploited by attackers to degrade network performance or even disrupt network avai...

The development of Internet of Things (IoT) dramatically facilitates the integration of computing systems with the physical world. However, as IoT devices are more easily to compromise than desktop computers, cybercriminals have founded IoT-based botnets to launch DDoS attacks with unprecedented traffic volume. To mitigate the damages associated wi...

Protocol fuzzers are widely used for finding vulnerabilities and security bugs in the program. The main techniques used by protocol fuzzers can be divided into 2 categories: generation-based and mutation-based fuzzing. The generation-based fuzzing generates data messages using an official specification (i.e., grammar), while the mutation-based fuzz...

Web scan is one of the most common network attacks on the Internet, in which an adversary probes one or more websites to discover exploitable information in order to perform further cyber attacks. For a coordinated web scan, an adversary controls multiple sources to achieve a large-scale scanning as well as detection evasion. In this paper, a novel...

Random domains are widely used in present network environment. In the benign services, they have been utilized as disposable domains to transfer one-time signals. However, in recent, based on the convenience of random domains, some cybercriminals have utilized them to launch DNS water torture attack, a kind of DDoS attack targeting on the authorita...

Domain Generation Algorithm (DGA) has been outfitted by various malware families to extend the resistance to the blacklist-based techniques. A lot of previous approaches have been developed to detect the DGA-based malware based on the lexical property of the random generated domains. Unfortunately, attackers can adjust their DGAs to produce domains...

The panoramic images are widely used in many applications. Saliency detection is an important task for panoramic image processing. Traditional saliency detection algorithms that are originally designed for conventional flat-2D images are not efficient for panoramic images due to their particular viewing way. Based on this consideration, we propose...

Under the multitask environment, on the one hand, the demand for spatio-temporal correlation of access control models continues to increase, on the other hand, multitask simultaneous authorization causes resource competition in the task execution phase, which may leads to the failure of task execution due to lack of some resouces. This paper propos...

With the arrival of Industry 4.0, more and more industrial control systems are connected with the outside world, which brings tremendous convenience to industrial production and control, and also introduces many potential security hazards. After a large number of attack cases analysis, we found that attacks in SCADA systems can be divided into inte...

Traffic aggregation has been considered as an effective approach to improve the radio resource utilization for M2M communication in LTE-A and beyond networks. In the LTE-A specification, the Relay Node (RN) is recommended to aggregate uplink M2M small-sized packets. However, the delay brought by the packets aggregation is inevitably increased, whic...

Attacks in industrial control systems vary widely and are influenced by many factors, including the intent of the attacker, the capabilities of the attacker, the sophistication of the attacking techniques, and his familiarity with the industrial control systems and industrial processes. Attacks against industrial control systems are not a simple ne...

The increasing use of common computer technologies in SCADA systems has brought convenience and at the same time introduced unprecedented security threats. SCADA system is becoming an attacker's target day by day. As attackers gain more understanding of industrial control systems, attacker's attack methods are more and more sophisticated, and the a...

SCADA (Supervisory Control and Data Acquisition) systems play a significant role in ICS (Industrial Control System). Safety, security and stable operation is crucial to the SCADA system. However, SCADA system currently faces many security threats, of which the most harmful is the APT attack. This paper analyzes the attack surface and its own vulner...

Attacks that use sophisticated and complex methods in-creased recently, aiming to infiltrate the Supervisory Control and Data Acquisition (SCADA) system and stay undetected. Therefore, attackers often get access to authorized permissions of SCADA and bring catastrophic damages by sending ‘legitimate’ control commands. Furthermore, insiders may also...

This paper analyzes the traffic of a current LTE network in China and investigates the joint optimization of content object caching and scheduling for in-radio access network (RAN) caches. Cooperative caching has been well recognized as a way of unleashing the ultimate potential of in-RAN caches, yet its feasibility is still unexplored. Moreover, c...

Moving Target Defense(MTD) provides a promising solution to reduce the chance of weakness exposure by constantly changing the target's attack surface. Though lots of MTD technologies have been researched to defend network attacks, there is little systematic study on security assessment of MTD. This paper proposes a novel method to quantify the secu...

On the basis of analyzing the logical structure of cyber physical system for power grid (GCPS), we proposed a modeling method of GCPS using the set theory. By studying the coupling principle of power flow and information flow, the GCPS was defined as the interconnection between the set of physical objects and the set of information objects. Then, t...

DNS protocol has been used by many malwares for command-and-control (C&C). To improve the resiliency of C&C communication, Domain Generation Algorithm (DGA) has been utilized by recent malwares such as Locky, Conficker and Zeus. Many detection systems have been introduced for DGA-based botnets detection. However, such botnets detection approaches s...