Malware detection techniques achieve great success with deeper insight into the semantics of malware. Among existing techniques, function call graph (FCG) based methods promising performance due to their prominent representations malware's functionalities. Meanwhile, recent adversarial attacks not only perturb feature vectors deceive classifiers (i.e., feature-space attacks) but also investigate how generate real evasive malware problem-space attacks). However, are limited inconsistent...

Graph-based Anomaly Detection (GAD) is becoming prevalent due to the powerful representation abilities of graphs as well recent advances in graph mining techniques. These GAD tools, however, expose a new attacking surface, ironically their unique advantage being able exploit relations among data. That is, attackers now can manipulate those (i.e., structure graph) allow some target nodes evade detection. In this paper, we vulnerability by designing type targeted structural poisoning attacks...

Signed social networks are widely used to model the trust relationships among online users in security-sensitive systems such as cryptocurrency trading platforms, where prediction plays a critical role. In this paper, we investigate how attackers could mislead by secretly manipulating signed networks. To end, first design effective poisoning attacks against representative models. The formulated hard bi-level optimization problems, for which propose several efficient approximation solutions....

Clinically deployed deep brain stimulation (DBS) for the treatment of Parkinson's disease operates in an open loop with fixed parameters, and this may result high energy consumption suboptimal therapy. The objective manuscript is to establish, through simulation a computational model, closed-loop control system that can automatically adjust parameters recover normal activity model neurons. Exaggerated beta band recognized as hallmark neurons globus pallidus internus (GPi) was used feedback...

The Fairness and Goodness Algorithm (FGA) is a widely used trust system in signed directed networks. However, attackers can manipulate scores on FGA by launching indirect Sybil attacks exploiting strong ties. In this work, we propose novel attack method vicinage-attack that formulates the problem as combination optimization for mining candidate attacking edges. Our constructs perturbation spaces infers existence of polymorphic To evaluate vicinage-attack, compare it against several baselines...

In this paper, we present a computational analysis of the problem attacking sign prediction, whereby aim attacker (a network member) is to hide from defender (an analyst) signs target set links by removing some other, non-target, links. The turns out be NP-hard if either local or global similarity measures are used for prediction. We propose heuristic algorithm and test its effectiveness on several real-life synthetic datasets.

Graph Neural Networks (GNNs) have become widely used in the field of graph mining. However, these networks are vulnerable to structural perturbations. While many research efforts focused on analyzing vulnerability through poisoning attacks, we identified an inefficiency current attack losses. These losses steer strategy towards modifying edges targeting misclassified nodes or resilient nodes, resulting a waste adversarial perturbation. To address this issue, propose novel loss framework...

Despite the tremendous success of graph neural networks in learning relational data, it has been widely investigated that are vulnerable to structural attacks on homophilic graphs. Motivated by this, a surge robust models is crafted enhance adversarial robustness However, vulnerability based heterophilic graphs remains mystery us. To bridge this gap, paper, we start explore and theoretically prove update negative classification loss negatively correlated with pairwise similarities powered...

The success of graph neural networks stimulates the prosperity mining and corresponding downstream tasks including anomaly detection (GAD). However, it has been explored that those methods are vulnerable to structural manipulations on relational data. That is, attacker can maliciously perturb structures assist target nodes in evading detection. In this article, we explore vulnerability two typical GAD systems: unsupervised FeXtra-based supervised convolutional network (GCN)-based GAD....

Parkinson's disease (PD) is a degenerative disorder of central nervous system that endangers the olds' health seriously. The motor symptoms PD can be attributed to distorted relay reliability thalamus cortical sensorimotor input results from increase inhibitory internal segment globus pallidum (GPi). Based on this, we construct GPi-thalamocortical computational model generate normal and pathological firing patterns by varying GPi spike train input. A kind closed-loop deep brain stimulation...

The akinesia symptoms for Parkinson's disease (PD) is characterized by the abnormal beta (13-30Hz) oscillations in basal ganglia. Deep brain stimulation (DBS) can alleviate movement disorders targeting include thalamic ventralis intermedius nucleus, subthalamic and globus pallidus. Here, we construct a computational model, an inhibitory loop between striatum pallidus externa (GPe). We present firing pattern spectral characteristics of striatal neurons GPe simulated healthy Parkinsonian...

The robustness of recommender systems under node injection attacks has garnered significant attention. Recently, GraphRfi, a Graph-Neural-Network-based (GNN-based) system, was proposed and shown to effectively mitigate the impact injected fake users. However, we demonstrate that GraphRfi remains vulnerable due supervised nature its fraudster detection component, where obtaining clean labels is challenging in practice. In particular, propose powerful poisoning attack, MetaC, against both...

Abstract In the realm of network security, adversarial attacks pose a significant threat, prompting research into innovative strategies for enhancing both attack and defense mechanisms. This paper presents novel exploration enhancement on Fairness Goodness Algorithm (FGA) Review to Reviewer (REV2), focusing trust prediction within signed graphs. contrast traditional time-based dynamics, propagation in FGA REV2 is grounded iterative processes. Through meticulous analysis structures, this...

We propose and analyze an instantaneous frequency measurement system by using optical power monitoring technique with improved resolution. The primary component adopted in the proposal is a dual-polarization quadrature phase shift keying (DP-QPSK) modulator which used to modulate microwave signal that has designed time delay shifting. generated sent polarization beam splitter (PBS) DP-QPSK modulator. Owing complementary transmission nature of interference introduced PBS, information...