Search Results (56)

Since the end of Apartheid in 1994, South Africa has striven to address past discrimination against members of marginalized groups such as Africans, women and LGBTQ+ individuals. Sophisticated media legislation and a vibrant civil society forged in the struggle against Apartheid ensure limited discrimination in traditional media and relatively fringe online forums. However, subtle forms of undermining signal the persistent legacy of a colonial and patriarchal past. While incidents of online racism and sexism are relatively well documented, ableism deserves more attention. Despite growing scholarship on digital discrimination, a significant research gap remains in understanding how ableist microaggressions manifest online, particularly when intersecting with race and gender. As a result of established media tropes, microaggressions against people with disabilities are somewhat naturalized and reproduced on social media, yet their intersectional dimensions—especially targeting Black women with disabilities—remain underexplored. This paper addresses this gap through a focused case study of jokes targeting Black women with disabilities in one TikTok video and the approximately 700 comments. Considering (dis)ability’s intersections with race, gender, and socio-economic status, these media texts are subjected to a critical thematic analysis. The study also problematizes the methodological challenges associated with finding, identifying, and purposively selecting such content. The analysis reveals a set of historically and contextually rooted microaggressions expressed through humor, which, as a cultural expression, is inherently covert and thus hard to detect and regulate. This research contributes to understanding how intersectional ableism operates digitally and highlights the need for nuanced approaches to identifying subtle forms of discrimination in online spaces. Full article

Cyberattacks, especially Advanced Persistent Threats (APTs), have become more complex. These evolving threats challenge traditional defense systems, which struggle to counter long-lasting and covert attacks. Cybersecurity Knowledge Graphs (CKGs), enabled through the integration of multi-source CTI, introduce novel approaches for proactive defense. However, building CKGs faces challenges such as unclear terminology, overlapping entity relationships in attack chains, and differences in CTI across sources. To tackle these challenges, we propose the CyberKG framework, which improves entity recognition and relation extraction using a SecureBERT_Plus-BiLSTM-Attention-CRF joint architecture. Semantic features are captured using a domain-adapted SecureBERT_Plus model, while temporal dependencies are modeled through BiLSTM. Attention mechanisms highlight key cross-sentence relationships, while CRF incorporates ATT&CK rule constraints. Hierarchical clustering (HAC), based on contextual embeddings, facilitates dynamic entity disambiguation and semantic fusion. Experimental evaluations on the DNRTI and MalwareDB datasets demonstrate strong performance in extraction accuracy, entity normalization, and the resolution of overlapping relations. The constructed knowledge graph supports APT tracking, attack-chain provenance, proactive defense prediction. Full article

Borderline personality disorder (BPD) is associated with alterations in emotion processing. To date, no study has tested automatic emotion perception under conditions of unawareness of emotion stimuli. We administered a priming paradigm based on facial expressions and measured judgmental and gaze behavior during an evaluation task. A total of 31 patients with BPD and 31 non-patients (NPs) viewed a briefly shown emotional (angry, fearful, sad, or happy) or neutral face followed by a neutral facial expression (target). Areas of interest (AOI) were the eyes and the mouth. All participants included in our analysis were subjectively unaware of the emotional primes. Concerning evaluative ratings, no prime effects were observed. For early dwell time, a significant interaction between prime category and AOI was found. Both BPD patients and NPs dwelled longer on the eyes after the presentation of angry and fearful primes than of happy primes and dwelled longer on the mouth after the presentation of happy primes than of sad and neutral primes. Patients rated target faces more negatively. BPD patients, when compared to NPs, seem not to show alterations in automatic attention orienting due to covert facial emotions. Regardless of primes, individuals with BPD seem to be characterized by an increased negative interpretation of neutral facial expressions. Full article

Advanced Persistent Threat (APT) attacks pose a serious challenge to traditional detection methods. These methods often suffer from high false-alarm rates and limited accuracy due to the multi-stage and covert nature of APT attacks. In this paper, we propose TSE-APT, a time-series ensemble model that addresses these two limitations. It combines multiple machine-learning models, such as Random Forest (RF), Multi-Layer Perceptron (MLP), and Bidirectional Long Short-Term Memory Network (BiLSTM) models, to dynamically capture correlations between multiple stages of the attack process based on time-series features. It discovers hidden features through the integration of multiple machine-learning models to significantly improve the accuracy and robustness of APT detection. First, we extract a collection of dynamic time-series features such as traffic mean, flow duration, and flag frequency. We fuse them with static contextual features, including the port service matrix and protocol type distribution, to effectively capture the multi-stage behaviors of APT attacks. Then, we utilize an ensemble-learning model with a dynamic weight-allocation mechanism using a self-attention network to adaptively adjust the sub-model contribution. The experiments showed that using time-series feature fusion significantly enhanced the detection performance. The RF, MLP, and BiLSTM models achieved 96.7% accuracy, considerably enhancing recall and the false positive rate. The adaptive mechanism optimizes the model’s performance and reduces false-alarm rates. This study provides an analytical method for APT attack detection, considering both temporal dynamics and context static characteristics, and provides new ideas for security protection in complex networks. Full article

Linguistic steganography can be utilized to establish covert communication channels on social media platforms, thus facilitating the dissemination of illegal messages, seriously compromising cyberspace security. Synonym substitution-based linguistic steganography methods have garnered considerable attention due to their simplicity and strong imperceptibility. Existing linguistic steganalysis methods have not achieved excellent detection performance for the aforementioned type of linguistic steganography. In this paper, based on the idea of focusing on accumulated differences, we propose a two-stage synonym substitution-based linguistic steganalysis method that does not require a synonym database and can effectively detect texts with very low embedding rates. Experimental results demonstrate that this method achieves an average detection accuracy 2.4% higher than the comparative method. Full article

Advanced persistent threat (APT) attacks present significant challenges to cybersecurity due to their covert nature, high complexity, and ability to operate across multiple temporal and spatial scales. Existing detection techniques often struggle with issues like class imbalance, insufficient feature extraction, and the inability to capture complex attack dependencies. To address these limitations, we propose a dual-phase framework for APT detection, combining multi-feature-conditioned generative adversarial networks (MF-CGANs) for data reconstruction and a multi-scale convolution and channel attention-enhanced graph convolutional network (MC-GCN) for improved attack detection. The MF-CGAN model generates minority-class samples to resolve the class imbalance problem, while MC-GCN leverages advanced feature extraction and graph convolution to better model the intricate relationships within network traffic data. Experimental results show that the proposed framework achieves significant improvements over baseline models. Specifically, MC-GCN outperforms traditional CNN-based IDS models, with accuracy, precision, recall, and F1-score improvements ranging from 0.47% to 13.41%. The MC-GCN model achieves an accuracy of 99.87%, surpassing CNN (86.46%) and GCN (99.24%), while also exhibiting high precision (99.87%) and recall (99.88%). These results highlight the proposed model’s superior ability to handle class imbalance and capture complex attack behaviors, establishing it as a leading approach for APT detection. Full article

In the realm of natural language processing (NLP), text classification constitutes a task of paramount significance for large language models (LLMs). Nevertheless, extant methodologies predominantly depend on the output generated by the final layer of LLMs, thereby neglecting the wealth of information encapsulated within neurons residing in intermediate layers. To surmount this shortcoming, we introduce LENS (Linear Exploration and Neuron Selection), an innovative technique designed to identify and sparsely integrate salient neurons from intermediate layers via a process of linear exploration. Subsequently, these neurons are transmitted to downstream modules dedicated to text classification. This strategy effectively mitigates noise originating from non-pertinent neurons, thereby enhancing both the accuracy and computational efficiency of the model. The detection of telecommunication fraud text represents a formidable challenge within NLP, primarily attributed to its increasingly covert nature and the inherent limitations of current detection algorithms. In an effort to tackle the challenges of data scarcity and suboptimal classification accuracy, we have developed the LENS-RMHR (Linear Exploration and Neuron Selection with RoBERTa, Multi-head Mechanism, and Residual Connections) model, which extends the LENS framework. By incorporating RoBERTa, a multi-head attention mechanism, and residual connections, the LENS-RMHR model augments the feature representation capabilities and improves training efficiency. Utilizing the CCL2023 telecommunications fraud dataset as a foundation, we have constructed an expanded dataset encompassing eight distinct categories that encapsulate a diverse array of fraud types. Furthermore, a dual-loss function has been employed to bolster the model’s performance in multi-class classification scenarios. Experimental results reveal that LENS-RMHR demonstrates superior performance across multiple benchmark datasets, underscoring its extensive potential for application in the domains of text classification and telecommunications fraud detection. Full article

Pupil responses and optokinetic nystagmus (OKN) are known to vary with the brightness and direction of motion of attended stimuli, as well as gaze position. However, whether these processes are controlled by a common mechanism remains unclear. In this study, we investigated how OKN latency relates to pupil response latency under two conditions: gaze shifts (eye movement) and attention shifts (covert attention without eye movement). As a result, while OKN showed consistent temporal changes across both gaze and attention conditions, pupillary responses exhibited distinct patterns. Moreover, the results revealed no significant correlation between pupil latency and OKN latency in either condition. These findings suggest that, although OKN and pupillary responses are influenced by similar attentional processes, their underlying mechanisms may differ. Full article

As Unmanned Aerial Vehicle (UAV) technology advances, UAVs have attracted widespread attention across military and civilian fields due to their low cost and flexibility. In unknown environments, UAVs can significantly reduce the risk of casualties and improve the safety and covertness when performing missions. Reinforcement Learning allows agents to learn optimal policies through trials in the environment, enabling UAVs to respond autonomously according to the real-time conditions. Due to the limitation of the observation range of UAV sensors, UAV target search missions face the challenge of partial observation. Based on this, Partially Observable Deep Q-Network (PODQN), which is a DQN-based algorithm is proposed. The PODQN algorithm utilizes the Gated Recurrent Unit (GRU) to remember the past observation information. It integrates the target network and decomposes the action value for better evaluation. In addition, the artificial potential field is introduced to solve the potential collision problem. The simulation environment for UAV target search is constructed through the custom Markov Decision Process. By comparing the PODQN algorithm with random strategy, DQN, Double DQN, Dueling DQN, VDN, QMIX, it is demonstrated that the proposed PODQN algorithm has the best performance under different agent configurations. Full article

Telecom fraud has emerged as one of the most pressing challenges in the criminal field. With advancements in artificial intelligence, telecom fraud texts have become increasingly covert and deceptive. Existing prevention methods, such as mobile number tracking, detection, and traditional machine-learning-based text recognition, struggle in terms of their real-time performance in identifying telecom fraud. Additionally, the scarcity of Chinese telecom fraud text data has limited research in this area. In this paper, we propose a telecom fraud text detection model, RoBERTa-MHARC, which combines RoBERTa with a multi-head attention mechanism and residual connections. First, the model selects data categories from the CCL2023 telecom fraud dataset as basic samples and merges them with collected telecom fraud text data, creating a five-category dataset covering impersonation of customer service, impersonation of leadership acquaintances, loans, public security fraud, and normal text. During training, the model integrates a multi-head attention mechanism and enhances its training efficiency through residual connections. Finally, the model improves its multi-class classification accuracy by incorporating an inconsistency loss function alongside the cross-entropy loss. The experimental results demonstrate that our model performs well on multiple benchmark datasets, achieving an F1 score of 97.65 on the FBS dataset, 98.10 on our own dataset, and 93.69 on the news dataset. Full article

Mental fatigue is a variation in the psychophysiological state that subjects encounter during or after prolonged cognitive activity periods, affecting top-down attention and cognitive control. The present study aimed to investigate the effects of mental fatigue on attention in the context of the three attention networks according to the Posnerian model (alerting, orienting, and executive networks) by combining the Attentional Network Test (ANT) and event-related potentials technique. Thirty healthy subjects were enrolled in the study. A continuous arithmetic task lasting one hour induced mental fatigue, and EEG recordings were conducted before and after the task while subjects were performing the ANT. The efficiencies of three networks were comparable between groups, while RTs shortened only in the control group and the accuracy related to the alerting and conflict networks declined only after mental effort. Mental fatigue reduced N1 amplitude during alerting network engagement and p3 amplitude during orienting. It also reduced N2 and P3 amplitude during the conflict, particularly the incongruent target-locked response. These findings underscore the covert effects of mental fatigue on attention, suggesting that even in healthy young subjects, compensatory mechanisms may maintain adequate overt performances, but fatigue still has a detrimental effect on top-down attentional mechanisms. Full article

Computer simulation of continuous chaotic systems is usually performed using numerical methods. The discretization may introduce new properties into finite-difference models compared to their continuous prototypes and can therefore lead to new types of dynamical behavior exhibited by discrete chaotic systems. It is known that one can control the dynamics of a discrete system using a special class of integration methods. One of the applications of such a phenomenon is chaos-based communication systems, which have recently attracted attention due to their high covertness and broadband transmission capability. Proper modulation of chaotic carrier signals is one of the key problems in chaos-based communication system design. It is challenging to modulate and demodulate a chaotic signal in the same way as a conventional signal due to its noise-like shape and broadband characteristics. Therefore, the development of new modulation–demodulation techniques is of great interest in the field. One possible approach here is to use adaptive numerical integration, which allows control of the properties of the finite-difference chaotic model. In this study, we describe a novel modulation technique for chaos-based communication systems based on generalized explicit second-order Runge–Kutta methods. We use a specially designed test bench to evaluate the efficiency of the proposed modulation method and compare it with state-of-the-art solutions. Experimental results show that the proposed modulation technique outperforms the conventional parametric modulation method in both coverage and noise immunity. The obtained results can be efficiently applied to the design of advanced chaos-based communication systems as well as being used to improve existing architectures. Full article

As the application of cyber-physical systems (CPSs) becomes more and more widespread, its security is becoming a focus of attention. Currently, there has been much research on the security defense of the physical layer of the CPS. However, most of the research only focuses on one of the aspects, for example, attack detection, security state estimation, or recovery control. Obviously, the effectiveness of security defense targeting only one aspect is limited. Therefore, in this paper, a set of security defense processes is proposed for the case that a CPS containing multiple sensors is subject to three kinds of stealthy attacks (i.e., zero-dynamics attack, covert attack, and replay attack). Firstly, the existing attack detection method based on improved residuals is used to detect stealthy attacks. Secondly, based on the detection results, an optimal state estimation method based on improved Kalman filtering is proposed to estimate the actual state of the system. Then, based on the optimal state, internal model control (IMC) is introduced to complete the recovery control of the system. Finally, the proposed methods are integrated to give a complete security defense process, and the simulation is verified for three kinds of stealthy attacks. The simulation results show that the proposed methods are effective. Full article

Amid the incessant evolution of the Internet, an array of cybersecurity threats has surged at an unprecedented rate. A notable antagonist within this plethora of attacks is the SQL injection assault, a prevalent form of Internet attack that poses a significant threat to web applications. These attacks are characterized by their extensive variety, rapid mutation, covert nature, and the substantial damage they can inflict. Existing SQL injection detection methods, such as static and dynamic detection and command randomization, are principally rule-based and suffer from low accuracy, high false positive (FP) rates, and false negative (FN) rates. Contemporary machine learning research on SQL injection attack (SQLIA) detection primarily focuses on feature extraction. The effectiveness of detection is heavily reliant on the precision of feature extraction, leading to a deficiency in tackling more intricate SQLIA. To address these challenges, we propose a novel SQLIA detection approach harnessing the power of an enhanced TextCNN and LSTM. This method begins by vectorizing the samples in the corpus and then leverages an improved TextCNN to extract local features. It then employs a Bidirectional LSTM (Bi-LSTM) network to decipher the sequence information inherent in the samples. Given LSTM’s modest effectiveness for relatively long sequences, we further integrate an attention mechanism, reducing the distance between any two words in the sequence to one, thereby enhancing the model’s effectiveness. Moreover, pre-trained word vector features acquired via BERT for transfer learning are incorporated into the feature section. Comparative experimental results affirm the superiority of our deep learning-based SQLIA detection approach, as it effectively elevates the SQLIA recognition rate while reducing both FP and FN rates. Full article

The fast development and adoption of IoT technologies has been enabling their application into increasingly sensitive domains, such as Medical and Industrial IoT, in which safety and cyber-security are paramount. While the number of deployed IoT devices increases annually, they still present severe cyber-security vulnerabilities, becoming potential targets and entry points for further attacks. As these nodes become compromised, attackers aim to set up stealthy communication behaviours, to exfiltrate data or to orchestrate nodes in a cloaked fashion, and network timing covert channels are increasingly being used with such malicious intents. The IEEE 802.15.4 is one of the most pervasive protocols in IoT and a fundamental part of many communication infrastructures. Despite this fact, the possibility of setting up such covert communication techniques on this medium has received very little attention. We aim to analyse the performance and feasibility of such covert-channel implementations upon the IEEE 802.15.4 protocol, particularly upon the DSME behaviour, one of the most promising for large-scale time critical communications. This enables us to better understand the involved risk of such threats and help support the development of active cyber-security mechanisms to mitigate these threats, which, for now, we provide in the form of practical network setup recommendations. Full article