Biblio
Phishing attacks are the most common form of attacks that can happen over the internet. This method involves attackers attempting to collect data of a user without his/her consent through emails, URLs, and any other link that leads to a deceptive page where a user is persuaded to commit specific actions that can lead to the successful completion of an attack. These attacks can allow an attacker to collect vital information of the user that can often allow the attacker to impersonate the victim and get things done that only the victim should have been able to do, such as carry out transactions, or message someone else, or simply accessing the victim's data. Many studies have been carried out to discuss possible approaches to prevent such attacks. This research work includes three machine learning algorithms to predict any websites' phishing status. In the experimentation these models are trained using URL based features and attempted to prevent Zero-Day attacks by using proposed software proposal that differentiates the legitimate websites and phishing websites by analyzing the website's URL. From observations, the random forest classifier performed with a precision of 97%, a recall 99%, and F1 Score is 97%. Proposed model is fast and efficient as it only works based on the URL and it does not use other resources for analysis, as was the case for past studies.
Cybersecurity community is slowly leveraging Machine Learning (ML) to combat ever evolving threats. One of the biggest drivers for successful adoption of these models is how well domain experts and users are able to understand and trust their functionality. As these black-box models are being employed to make important predictions, the demand for transparency and explainability is increasing from the stakeholders.Explanations supporting the output of ML models are crucial in cyber security, where experts require far more information from the model than a simple binary output for their analysis. Recent approaches in the literature have focused on three different areas: (a) creating and improving explainability methods which help users better understand the internal workings of ML models and their outputs; (b) attacks on interpreters in white box setting; (c) defining the exact properties and metrics of the explanations generated by models. However, they have not covered, the security properties and threat models relevant to cybersecurity domain, and attacks on explainable models in black box settings.In this paper, we bridge this gap by proposing a taxonomy for Explainable Artificial Intelligence (XAI) methods, covering various security properties and threat models relevant to cyber security domain. We design a novel black box attack for analyzing the consistency, correctness and confidence security properties of gradient based XAI methods. We validate our proposed system on 3 security-relevant data-sets and models, and demonstrate that the method achieves attacker's goal of misleading both the classifier and explanation report and, only explainability method without affecting the classifier output. Our evaluation of the proposed approach shows promising results and can help in designing secure and robust XAI methods.
Artificial Intelligence systems have enabled significant benefits for users and society, but whilst the data for their feeding are always increasing, a side to privacy and security leaks is offered. The severe vulnerabilities to the right to privacy obliged governments to enact specific regulations to ensure privacy preservation in any kind of transaction involving sensitive information. In the case of digital and/or physical documents comprising sensitive information, the right to privacy can be preserved by data obfuscation procedures. The capability of recognizing sensitive information for obfuscation is typically entrusted to the experience of human experts, who are over-whelmed by the ever increasing amount of documents to process. Artificial intelligence could proficiently mitigate the effort of the human officers and speed up processes. Anyway, until enough knowledge won't be available in a machine readable format, automatic and effectively working systems can't be developed. In this work we propose a methodology for transferring and leveraging general knowledge across specific-domain tasks. We built, from scratch, specific-domain knowledge data sets, for training artificial intelligence models supporting human experts in privacy preserving tasks. We exploited a mixture of natural language processing techniques applied to unlabeled domain-specific documents corpora for automatically obtain labeled documents, where sensitive information are recognized and tagged. We performed preliminary tests just over 10.000 documents from the healthcare and justice domains. Human experts supported us during the validation. Results we obtained, estimated in terms of precision, recall and F1-score metrics across these two domains, were promising and encouraged us to further investigations.
Conflicts may arise at any time during military debriefing meetings, especially in high intensity deployed settings. When such conflicts arise, it takes time to get everyone back into a receptive state of mind so that they engage in reflective discussion rather than unproductive arguing. It has been proposed by some that the use of social robots equipped with social abilities such as emotion regulation through rapport building may help to deescalate these situations to facilitate critical operational decisions. However, in military settings, the same AI agent used in the pre-brief of a mission may not be the same one used in the debrief. The purpose of this study was to determine whether a brief rapport-building session with a social robot could create a connection between a human and a robot agent, and whether consistency in the embodiment of the robot agent was necessary for maintaining this connection once formed. We report the results of a pilot study conducted at the United States Air Force Academy which simulated a military mission (i.e., Gravity and Strike). Participants' connection with the agent, sense of trust, and overall likeability revealed that early rapport building can be beneficial for military missions.
Trust is an important characteristic of successful interactions between humans and agents in many scenarios. Self-driving scenarios are of particular relevance when discussing the issue of trust due to the high-risk nature of erroneous decisions being made. The present study aims to investigate decision-making and aspects of trust in a realistic driving scenario in which an autonomous agent provides guidance to humans. To this end, a simulated driving environment based on a college campus was developed and presented. An online and an in-person experiment were conducted to examine the impacts of mistakes made by the self-driving AI agent on participants’ decisions and trust. During the experiments, participants were asked to complete a series of driving tasks and make a sequence of decisions in a time-limited situation. Behavior analysis indicated a similar relative trend in the decisions across these two experiments. Survey results revealed that a mistake made by the self-driving AI agent at the beginning had a significant impact on participants’ trust. In addition, similar overall experience and feelings across the two experimental conditions were reported. The findings in this study add to our understanding of trust in human-robot interaction scenarios and provide valuable insights for future research work in the field of human-robot trust.
Advancements in the AI field unfold tremendous opportunities for society. Simultaneously, it becomes increasingly important to address emerging ramifications. Thereby, the focus is often set on ethical and safe design forestalling unintentional failures. However, cybersecurity-oriented approaches to AI safety additionally consider instantiations of intentional malice – including unethical malevolent AI design. Recently, an analogous emphasis on malicious actors has been expressed regarding security and safety for virtual reality (VR). In this vein, while the intersection of AI and VR (AIVR) offers a wide array of beneficial cross-fertilization possibilities, it is responsible to anticipate future malicious AIVR design from the onset on given the potential socio-psycho-technological impacts. For a simplified illustration, this paper analyzes the conceivable use case of Generative AI (here deepfake techniques) utilized for disinformation in immersive journalism. In our view, defenses against such future AIVR safety risks related to falsehood in immersive settings should be transdisciplinarily conceived from an immersive co-creation stance. As a first step, we motivate a cybersecurity-oriented procedure to generate defenses via immersive design fictions. Overall, there may be no panacea but updatable transdisciplinary tools including AIVR itself could be used to incrementally defend against malicious actors in AIVR.
With the development of Internet technology, the attacker gets more and more complex background knowledge, which makes the anonymous model susceptible to background attack. Although the differential privacy model can resist the background attack, it reduces the versatility of the data. In this paper, this paper proposes a differential privacy information publishing algorithm based on clustering anonymity. The algorithm uses the cluster anonymous algorithm based on KD tree to cluster the original data sets and gets anonymous tables by anonymous operation. Finally, the algorithm adds noise to the anonymous table to satisfy the definition of differential privacy. The algorithm is compared with the DCMDP (Density-Based Clustering Mechanism with Differential Privacy, DCMDP) algorithm under different privacy budgets. The experiments show that as the privacy budget increases, the algorithm reduces the information loss by about 80% of the published data.
Video Surveillance plays a pivotal role in today's world. The technologies have been advanced too much when artificial intelligence, machine learning and deep learning pitched into the system. Using above combinations, different systems are in place which helps to differentiate various suspicious behaviors from the live tracking of footages. The most unpredictable one is human behaviour and it is very difficult to find whether it is suspicious or normal. Deep learning approach is used to detect suspicious or normal activity in an academic environment, and which sends an alert message to the corresponding authority, in case of predicting a suspicious activity. Monitoring is often performed through consecutive frames which are extracted from the video. The entire framework is divided into two parts. In the first part, the features are computed from video frames and in second part, based on the obtained features classifier predict the class as suspicious or normal.
Big Data Platform provides business units with data platforms, data products and data services by integrating all data to fully analyze and exploit the intrinsic value of data. Data accessed by big data platforms may include many users' privacy and sensitive information, such as the user's hotel stay history, user payment information, etc., which is at risk of leakage. This paper first analyzes the risks of data leakage, then introduces in detail the theoretical basis and common methods of data desensitization technology, and finally puts forward a set of effective market subject credit supervision application based on asccii, which is committed to solving the problems of insufficient breadth and depth of data utilization for enterprises involved, the problems of lagging regulatory laws and standards, the problems of separating credit construction and market supervision business, and the credit constraints of data governance.