Competitive Attack Research Articles

Effective and Imperceptible Adversarial Textual Attack Via Multi-objectivization

The field of adversarial textual attack has significantly grown over the past few years, where the commonly considered objective is to craft adversarial examples (AEs) that can successfully fool the target model. However, the imperceptibility of attacks, which is also essential for practical attackers, is often left out by previous studies. In consequence, the crafted AEs tend to have obvious structural and semantic differences from the original human-written text, making them easily perceptible. In this work, we advocate leveraging multi-objectivization to address such an issue. Specifically, we reformulate the problem of crafting AEs as a multi-objective optimization problem, where the attack imperceptibility is considered as an auxiliary objective. Then, we propose a simple yet effective evolutionary algorithm, dubbed HydraText, to solve this problem. HydraText can be effectively applied to both score-based and decision-based attack settings. Exhaustive experiments involving 44,237 instances demonstrate that HydraText consistently achieves competitive attack success rates and better attack imperceptibility than the recently proposed attack approaches. A human evaluation study also shows that the AEs crafted by HydraText are more indistinguishable from human-written text. Finally, these AEs exhibit good transferability and can bring notable robustness improvement to the target model by adversarial training.

Invisible Backdoor Attack against 3D Point Cloud Classifier in Graph Spectral Domain

3D point cloud has been wildly used in security crucial domains, such as self-driving and 3D face recognition. Backdoor attack is a serious threat that usually destroy Deep Neural Networks (DNN) in the training stage. Though a few 3D backdoor attacks are designed to achieve guaranteed attack efficiency, their deformation will alarm human inspection. To obtain invisible backdoored point cloud, this paper proposes a novel 3D backdoor attack, named IBAPC, which generates backdoor trigger in the graph spectral domain. The effectiveness is grounded by the advantage of graph spectral signal that it can induce both global structure and local points to be responsible for the caused deformation in spatial domain. In detail, a new backdoor implanting function is proposed whose aim is to transform point cloud to graph spectral signal for conducting backdoor trigger. Then, we design a backdoor training procedure which updates the parameter of backdoor implanting function and victim 3D DNN alternately. Finally, the backdoored 3D DNN and its associated backdoor implanting function is obtained by finishing the backdoor training procedure. Experiment results suggest that IBAPC achieves SOTA attack stealthiness from three aspects including objective distance measurement, subjective human evaluation, graph spectral signal residual. At the same time, it obtains competitive attack efficiency. The code is available at https://github.com/f-lk/IBAPC.

The Role of Uig Linkages in Improving Competitive Advantage: A Case of the Clothing, Textiles, Leather, and Footwear Industry of South Africa

The South African clothing, textiles, leather, and footwear (CTLF) sector has been under competitive attack, mainly from Asian countries. The many studies undertaken so far have not focused on the role of research and development linkages among university, industry, and government (UIG), and hence, this study was undertaken. This study uses mixed methods, interviews of managers and directors from UIG actors, and a survey of CTLF firms to understand the role of UIG linkages in improving competitive advantage. The data was thematically and descriptively analysed. The findings highlight that most CTLF firms were not involved in UIG research and development linkages. As a result, CTLF firms have not utilised researchers available in the universities nor research-related industry and government funding that supports innovation and technological improvements. For the CTLF to improve its competitive advantage, its development strategies must include being engaged in UIG research and development linkages. 
 
 Received: 28 September 2023 / Accepted: 3 February 2024 / Published: 5 March 2024

Solving Degree Bounds for Iterated Polynomial Systems

For Arithmetization-Oriented ciphers and hash functions Gröbner basis attacks are generally considered as the most competitive attack vector. Unfortunately, the complexity of Gröbner basis algorithms is only understood for special cases, and it is needless to say that these cases do not apply to most cryptographic polynomial systems. Therefore, cryptographers have to resort to experiments, extrapolations and hypotheses to assess the security of their designs. One established measure to quantify the complexity of linear algebra-based Gröbner basis algorithms is the so-called solving degree. Caminata & Gorla revealed that under a certain genericity condition on a polynomial system the solving degree is always upper bounded by the Castelnuovo-Mumford regularity and henceforth by the Macaulay bound, which only takes the degrees and number of variables of the input polynomials into account. In this paper we extend their framework to iterated polynomial systems, the standard polynomial model for symmetric ciphers and hash functions. In particular, we prove solving degree bounds for various attacks on MiMC, Feistel-MiMC, Feistel-MiMC-Hash, Hades and GMiMC. Our bounds fall in line with the hypothesized complexity of Gröbner basis attacks on these designs, and to the best of our knowledge this is the first time that a mathematical proof for these complexities is provided. Moreover, by studying polynomials with degree falls we can prove lower bounds on the Castelnuovo-Mumford regularity for attacks on MiMC, Feistel-MiMC and Feistel-MiMCHash provided that only a few solutions of the corresponding iterated polynomial system originate from the base field. Hence, regularity-based solving degree estimations can never surpass a certain threshold, a desirable property for cryptographic polynomial systems.

External Disruption Following CEO Succession: How Human and Social Capital Changes May Attract Competitive Attacks

External Disruption Following CEO Succession: How Human and Social Capital Changes May Attract Competitive Attacks

Система верификации личности по изображению лица в защищенном режиме на основе искусственных нейронных сетей

The aim of the work is to develop a system for verifying subjects by face based on a neural network model that is executed in a protected mode. Protected mode means that the identity verification system is highly resistant to destructive influences, such as competitive attacks, and allows storing and processing biometric data without compromising it. The system is based on a biometrics to code converter trained according to GOST R 52633.5, which allows you to associate the subject’s facial biometrics image with its cryptographic key or long password, which can later be used for authentication, and deep convolutional neural networks. For face detection in the image, the MTCNN artificial neural network architecture was used, and several neural network architectures were tested for feature extraction: InceptionResnet, Facenet512, VGG-Face and OpenFace. The best results were shown by the InceptionResnet neural network. When evaluating the effectiveness and testing the reliability of the proposed system on a special dataset of faces collected under different lighting conditions in a room, it was possible to achieve a relatively low value of equal probability of errors of the first and second kind (EER = 0.0146 with a key length of 278 bits), which confirms the effectiveness of the considered approach to building face verification systems.

Unsupervised cycle-consistent adversarial attacks for visual object tracking

Adversarial attacks on visual object tracking have been gaining momentum in the quest to assess and enhance the robustness and security of object tracking models. Most adversarial attack strategies for object tracking rely on fully supervised attacks; that is, they necessitate the use of labels. However, it is not feasible to apply these labeled video object tracking techniques to all attacks. As such, this paper presents an unsupervised attack methodology against visual object tracking models. The approach employs the cycle consistency principle of object tracking models to maximize the inconsistency between forward and backward tracking, thereby providing effective countermeasures. Additionally, this paper introduces a contextual attack method, leveraging the information from the attack object’s region and its surrounding contextual regions. This strategy attacks the object region and its surrounding context regions simultaneously, aiming to decrease its response score to the attack. The proposed attack method is assessed across various types of deep learning-based object trackers. Experimental results on multiple benchmarks reveal that the suggested method yields competitive attack outcomes.

Determinants of Political Responses to A Competitive Attack

Determinants of Political Responses to A Competitive Attack

F− competitive attack decomposing parasitic product Al(OH)3 of hydrogel-based Al-air battery

Wearable Al-air batteries are regarded as the potential power systems for flexible electronics due to the ultra-high capacity and energy density of Al-based materials. However, the battery failure caused by the accumulation of parasitic product Al(OH)3 upon the anode surface has hindered the commercialization. Herein, we report a polyacrylic acid hydrogel integrating KF and KOH (F@PAA), which decompose Al(OH)3 for ameliorating discharge performance of wearable Al-air battery. The ions channels upon the anode surface are dredged by a competitive attack of F− on Al-O, thus improving the battery durability. The results show that the binding of Al3+ with F− is more stable than that with O2−. The formed complex AlF63− corrodes the passivation layer, and then ensures the continuous anodic oxidation. When 1.0 M F− is introduced into F@PAA hydrogel, the effect of byproduct decomposition and battery discharge are optimal. Hence, A wearable Al-air battery using the proposed hydrogel achieves a maximum power density of 58.28 mW/cm2. A high capacity of 2199.10 mAh/g and anode efficiency of 73.80% for the battery can be obtained at 10 mA/cm2. Moreover, the key performance of the battery is improved by up to 104.08%, developing interface cleaning technology in wearable Al-air batteries.

Revisiting mean reversion in profitability and earnings: evidence from India (2007–2020)

PurposeThe purpose of this article is to revisit the mean reversion in profitability and earnings among Indian-listed firms, based on the idea that changes in profitability and earnings are somewhat predictable.Design/methodology/approachThe study used a sample of 445 Bombay Stock Exchange (BSE)-listed companies and 309 companies from the manufacturing sector in India for the period from 2007 to 2020. The study employed cross-sectional regressions. Both linear and non-linear Partial Adjustment Models (PAM) were used to forecast profitability and earnings.FindingsThe study revealed that profitability and earnings mean revert for both the BSE-listed companies and the manufacturing sector companies from 2007 to 2012. However, for the years from 2013 to 2020, it was found that there is no significant evidence of mean reversion in both the BSE-listed companies or the manufacturing sector companies.Practical implicationsThe findings have larger implications for security analysts who forecast future stabilisation or recovery of historically high or low growth rates. Investors and analysts would benefit from having a better understanding of how competitive attacks affect profitability as well as how the overall economic growth of a country affects earnings and valuations.Originality/valueMost of the empirical research in India has focused on mean reversion in stock prices or stock returns. The present study looked at the mean reversion of profitability and earnings in Indian firms.

Vulnerability in Transition Period: Newly appointed CEO Dismissals and Rivals’ Competitive Actions

Dismissal of newly appointed CEOs has become increasingly prevalent and attracted significant scholars’ and practitioners’ attention. Prior studies on this topic have mainly focused on its implication for the focal firm. This study extends the extant research by investigating how a focal firm’s dismissal of its newly appointed CEO may influence its rivals’ competitive actions. Building on CEO dismissal and competitive dynamics literature, we argue that as focal firm may experience organizational disruption from dismissing its newly appointed CEO, its rivals may seize opportunities in the vulnerable transition period and benefit from initiating more intense competitive actions toward the focal firm. Using data on dismissals of newly appointed CEO and rivals’ competitive attacks in public firms listed in United States from 2000 to 2019, we find that rival firms initiate more intense competitive actions upon the announcement of the focal firm’s dismissal of its newly appointed CEO. Moreover, the main effect would be moderated by the degree of competition intensity in the focal-competitor dyad level, which are captured by geographic market overlap and product similarity. We find that geographic market overlap and product similarity between focal and competitor firms both will strengthen our main effect. These results suggest that closer competitors to the focal firm benefit more from the focal firm’s disruption. This study contributes to the CEO dismissal literature by demonstrating the impact of CEO dismissal beyond the focal firm.

CEO Succession and Competitive Attacks from Rivals

CEO Succession and Competitive Attacks from Rivals

Situational Experience, Expertise, and the Janus-face of Forbearance

We study the drivers and performance outcomes of strategic forbearance – a purposeful decision not to respond to a rival’s competitive attack. Leveraging theories of skill development, we argue that forbearance can be strategically beneficial or not depending on competitors’ situational experience and level of expertise. We find that competitors with extensive experience in dealing with specific competitive situations are more likely to choose forbearance over response; however they benefit from forbearance only when they are skillful enough to create long-term positional advantage; otherwise, forbearance leads to inferior performance. Additionally, expert competitors with the highest skills favor response over forbearance, but benefit from both forbearance and response, whereas forbearance by those of limited expertise hurts performance. We operationalize forbearance as gambit decline in chess and test our propositions on a sample of 3,515 chess players making 236,767 decisions to accept or decline a gambit offer.

Why do High-Status Firms Become Targets of Competitive Attacks? An Examination of Two Explanations

Competitive dynamics research has often used the AMC framework—awareness, motivation, and capability—when theorizing how different factors influence firms’ proclivities to launch competitive moves against other firms. However, this stream of research has mostly overlooked the role of social evaluations in determining which firms are targeted. We study how a specific form of social evaluation, relative status, impacts motivation to act competitively against other firms. Research is divided on how relative status affects competitive moves. Some research suggests that awareness mediates the relationship between relative status and motivation. Other research suggests that relative status moderates the relationship between awareness and motivation. It is important to resolve the relationship between relative status, awareness, and motivation because it can help firms take actions to lower the likelihood that they will become targets of competitive attacks by other firms. Teasing out the effects of the two competing explanations has been difficult, however, because relative status, attention, and motivation are typically not measured directly. We test both mechanisms in a unique dataset of food trucks in a mid-sized city in the U.S. that allows us to overcome this limitation. We find that high-status food trucks elicit greater competitive motivation from other food trucks primarily due to the greater attention the high-status food trucks garner. Further, we find no support for a moderation effect. Our study helps understand the interplay between awareness and motivation in the context of social evaluations.

FE-DaST: Fast and effective data-free substitute training for black-box adversarial attacks

Deep learning models have shown their advantages in computer vision, e.g., image classification, whereas they are well-known to be susceptible to imperceptible perturbations of input images which are called adversarial attacks. Recently proposed data-free substitute training (DaST), an adversarial framework based on a multi-branch generator where each branch generated images of the corresponding class to balance synthetic images, trained surrogate models without the requirement of any real image for transfer-based black-box adversarial attacks. However, this multi-branch framework was too redundant to converge quickly and was limited to datasets of a few categories. In this paper, we propose a simpler adversarial framework based on a single-branch generator to train substitute models fast and effectively, named FE-DaST. More specifically, we adopt a single-branch deep convolutional generator with an information entropy loss to stimulate the generation of balanced images, promote the similarity between substitute models and target models, and further enhance the strength of the transfer-based attack. Despite its simplicity, experimental results demonstrate the superiority of our proposed FE-DaST over DaST in terms of computational loads, similarities between surrogate models and target models, and attack success rates of transferable adversarial examples on MNIST and CIFAR-10 datasets. For CIFAR-100 and Tiny-ImageNet datasets where DaST is not available, our FE-DaST also achieves competitive attack success rates compared with pre-trained models which are trained with realistic training images. Furthermore, the attack performance of FE-DaST outperforms other state-of-the-art substitute training methods on the four datasets.

Impact of Corporate Social Responsibility on Competitive Advantage: A Case of Newmont Ghana Limited

Corporate Social Responsibility has taken part in the operations of businesses, but there is a need for research on why CSR is and how advantageous it is for a firm in a competitive market. The paper investigated the use of CSR as a stronghold for firms in a competitive environment using Newmont Ghana Limited as a case study. The study further seeks to identify the impact of Newmont Ghana’s CSR activities and its impact on both the firm’s output and reputational image. These objectives are explored through the evaluation of the effect of Newmont Ghana’s CSR activities. Using a purposive and convenient sampling method, 120 questionnaires, both open and close-ended questions, were distributed to the employees of Newmont Ghana across all levels. The questionnaires were collected and analyzed. The outcome showed that corporate social responsibility gives Newmont Ghana an advantage in positioning and diminishing of competitive attack in the market. Also, the results showed that Newmont Ghana’s CSR activities have a positive impact on its reputation at both the local and national level. Also, CSR activities had a positive impact on its output as it draws in more investors and makes them perform better. The study recommended that firms should integrate CSR into their basic operations and to educate their employees on the firm’s social responsibilities, the general public should be educated on the different duties of all other actors, and the topic should be expanded to include all firms in the mining industry in Ghana and how they benefit from CSR.

Adversarial Attack on Large Scale Graph

Recent studies have shown that graph neural networks (GNNs) are vulnerable against perturbations due to lack of robustness and can therefore be easily fooled. Currently, most works on attacking GNNs are mainly using gradient information to guide the attack and achieve outstanding performance. However, the high complexity of time and space makes them unmanageable for large scale graphs and becomes the major bottleneck that prevents the practical usage. We argue that the main reason is that they have to use the whole graph for attacks, resulting in the increasing time and space complexity as the data scale grows. In this work, we propose an efficient Simplified Gradient-based Attack (SGA) method to bridge this gap. SGA can cause the GNNs to misclassify specific target nodes through a multi-stage attack framework, which needs only a much smaller subgraph. In addition, we present a practical metric named Degree Assortativity Change (DAC) to measure the impacts of adversarial attacks on graph data. We evaluate our attack method on four real-world graph networks by attacking several commonly used GNNs. The experimental results demonstrate that SGA can achieve significant time and memory efficiency improvements while maintaining competitive attack performance compared to state-of-art attack techniques.

Forbearance: Strategic Nonresponse to Competitive Attacks

Evidence suggests that firms often do not respond to competitive attacks even when they are aware of an attack and have the capability to counter it. We believe that this is because they engage in ...

Second‐order effects of CEO characteristics: How rivals' perceptions of CEOs as submissive and provocative precipitate competitive attacks

Research SummaryIntegrating victimization into competitive dynamics and upper echelons theorizing, we develop and test theory articulating how rivals' perceptions of a CEO precipitate attacks on the CEO's firm. Rather than treating CEOs' characteristics solely as perpetrating action (a first‐order effect, like research integrating upper echelons into competitive dynamics), we argue firms with CEOs possessing characteristics perceived as more submissive or more provocative are subject to more competitive actions directed toward their firms (a second‐order effect, like victimization research). Empirical analyses of a sample of Fortune 500 CEOs supports our theorizing while interviews of executives corroborate our premise as well. Our framework offers a more complete and socialized understanding of CEOs' roles in competitive dynamics, providing both theoretical and practical insights as well as future research avenues.Managerial SummaryWe articulate how CEOs possessing certain psychological, behavioral, and social characteristics may unknowingly precipitate competitive attacks on their firms. Our explanation integrates insights from victimology which explain how individuals are subject to more attacks if they possess characteristics others perceive as more submissive or more provocative. While prior research articulates that CEOs' characteristics affect decisions such as attacking rivals, integrating theories of victimization into this line of inquiry paints a more socialized view of why firms may be subject to competitive attacks as well. The logic and evidence we provide advances theoretical explanations of firms' competitive behaviors and executives' roles therein. At the same time, providing knowledge about how CEO characteristics precipitate competitive actions toward their firms can aid in prevention and intervention strategies.

When Forbearance is Preferable to Counteraction and Words: A Model of Competitive Response

The author introduces the construct of forbearance response and develops a theoretical model for predicting when defenders will respond to competitive attacks with forbearance, counteraction, or wo...