Biblio

In this short paper, we survey some work at the intersection of Artificial Intelligence (AI) and security that are based on game theoretic considerations, and particularly focus on the author's (our) contribution in these areas. One half of this paper focuses on applications of game theoretic and learning reasoning for addressing security applications such as in public safety and wildlife conservation. In the second half, we present recent work that attacks the learning components of these works, leading to sub-optimal defense allocation. We finally end by pointing to issues and potential research problems that can arise due to data quality in the real world.

Mortality from lung cancer has ranked high among cancers for many years. Early detection of lung cancer is critical for disease prevention, cure, and mortality rate reduction. Many existing detection methods on lung nodules can achieve high sensitivity but meanwhile introduce an excessive number of false-positive proposals, which is clinically unpractical. In this paper, we propose the multi-head detection and spatial attention network, shortly MHSnet, to address this crucial false-positive issue. Specifically, we first introduce multi-head detectors and skip connections to capture multi-scale features so as to customize for the variety of nodules in sizes, shapes, and types. Then, inspired by how experienced clinicians screen CT images, we implemented a spatial attention module to enable the network to focus on different regions, which can successfully distinguish nodules from noisy tissues. Finally, we designed a lightweight but effective false-positive reduction module to cut down the number of false-positive proposals, without any constraints on the front network. Compared with the state-of-the-art models, our extensive experimental results show the superiority of this MHSnet not only in the average FROC but also in the false discovery rate (2.64% improvement for the average FROC, 6.39% decrease for the false discovery rate). The false-positive reduction module takes a further step to decrease the false discovery rate by 14.29%, indicating its very promising utility of reducing distracted proposals for the downstream tasks relied on detection results.

Analyzing reflectionally symmetric features inside an image is one of the important processes for recognizing the peculiar appearance of natural and man-made objects, biological patterns, etc. In this work, we will point out an efficient detector of reflectionally symmetric shapes by addressing a class of projection-based signatures that are structured by a generalized \textbackslashmathcalR\_fm-transform model. To this end, we will firstly prove the \textbackslashmathcalR\_fmˆ-transform in accordance with reflectional symmetry detection. Then different corresponding \textbackslashmathcalR\_fm-signatures of binary shapes are evaluated in order to determine which the corresponding exponentiation of the \textbackslashmathcalR\_fm-transform is the best for the detection. Experimental results of detecting on single/compound contour-based shapes have validated that the exponentiation of 10 is the most discriminatory, with over 2.7% better performance on the multiple-axis shapes in comparison with the conventional one. Additionally, the proposed detector also outperforms most of other existing methods. This finding should be recommended for applications in practice.

Fraud mechanisms have evolved from isolated actions performed by single individuals to complex criminal networks. This paper aims to contribute to the identification of potentially relevant nodes in fraud networks. Whilst traditional methods for fraud detection rely on identifying abnormal patterns, this paper proposes STARBRIDGE: a new linear and scalable, ranked out, parameter free method to identify fraudulent nodes and rings based on Bridging, Influence and Control metrics. This is applied to the telecommunications domain where fraudulent nodes form a star-bridge-star pattern. Over 75% of nodes involved in fraud denote control, bridging centrality and doubled the influence scores, when compared to non-fraudulent nodes in the same role, stars and bridges being chief positions.

In this study, the parameters of the Jiles-Atherton (JA) model are investigated to determine suitable solution candidates for hysteresis models of a piezoelectric actuator (PEA). The methodology of this study is to perform Monte Carlo experiments on the JA model by randomly selecting parameters that generate hysteresis curves. The solution space is then restrained such that their normalized area and remnant displacements are comparable to those of the PEA. The data resulting from these Monte Carlo simulations show trends in the parameter space that can be used to further restrain parameter selection windows to find suitable JA parameters to model PEAs. In particular, the results show that selection of the reversibility coefficient and the pinning factor strongly affect both of the hysteresis characteristics studied. A large density of solutions is found in certain parameter distributions for both the area and the remnant displacement, but the remnant displacement generates the densest distributions. These results can be used to more effectively find suitable hysteresis models for modeling purposes.

Public Key Infrastructure (PKI) as a techno-policy ecosystem for establishing electronic trust has survived for several decades and evolved as the de-facto model for centralized trust in electronic transactions. In this paper, we study the PKI ecosystem that are prevailing in the South Asian and Oceanic countries and brief them. We also look at how PKI has coped up with the rapid technological changes and how policies have been realigned or formulated to strengthen the PKI ecosystem in these countries.

Recently, research on AI-based network intrusion detection has been actively conducted. In previous studies, the machine learning models such as SVM (Support Vector Machine) and RF (Random Forest) showed consistently high performance, whereas the NB (Naïve Bayes) showed various performances with large deviations. In the paper, after analyzing the cause of the NB models showing various performances addressed in the several studies, we measured the performance of the Gaussian NB model according to the smoothing factor that is closely related to these causes. Furthermore, we compared the performance of the Gaussian NB model with that of the other models as a zero-day attack detection system. As a result of the experiment, the accuracy was 38.80% and 87.99% in case that the smoothing factor is 0 and default respectively, and the highest accuracy was 94.53% in case that the smoothing factor is 1e-01. In the experiment, we used only some types of the attack data in the NSL-KDD dataset. The experiments showed the applicability of the Gaussian NB model as a zero-day attack detection system in the future. In addition, it is clarified that the smoothing factor of the Gaussian NB model determines the shape of gaussian distribution that is related to the likelihood.

The rapid shift towards smart cities, particularly in the era of pandemics, necessitates the employment of e-learning, remote learning systems, and hybrid models. Building adaptive and personalized education becomes a requirement to mitigate the downsides of distant learning while maintaining high levels of achievement. Explainable artificial intelligence (XAI), machine learning (ML), and the internet of behaviour (IoB) are just a few of the technologies that are helping to shape the future of smart education in the age of smart cities through Customization and personalization. This study presents a paradigm for smart education based on the integration of XAI and IoB technologies. The research uses data acquired on students' behaviours to determine whether or not the current education systems respond appropriately to learners' requirements. Despite the existence of sophisticated education systems, they have not yet reached the degree of development that allows them to be tailored to learners' cognitive needs and support them in the absence of face-to-face instruction. The study collected data on 41 learner's behaviours in response to academic activities and assessed whether the running systems were able to capture such behaviours and respond appropriately or not; the study used evaluation methods that demonstrated that there is a change in students' academic progression concerning monitoring using IoT/IoB to enable a relative response to support their progression.

Decentralized Identifiers (DIDs) and Self-Sovereign Identity (SSI) are emerging decentralized identity solutions. DIDs allow legal entities like organizations to create and fully control their identifiers while building the necessary infrastructure for SSI, enabling entities like persons, organizations, or machines to fully control and own their digital identities without the involvement of an intermediate central authority. DIDs are identifiers that are used to reference entities unambiguously and, together with DID Documents stored in a verifiable data registry, establish a new, decentralized public-key infrastructure. An SSI-based digital identity may be composed of many different claims certified by an issuer. Examples are the identity holder’s name, age, gender, university degree, driving license, or other attributes. What makes SSI unique compared to other identity management solutions is that the users keep their digital identities in storage of their choice and thus determine their distribution and processing.With this privacy-by-design approach, the emergence of DIDs and SSI can shape the architecture of the future Internet and its applications, which will impact the future of mobile networks. While 5G networks are currently being rolled out, a discussion about the new capabilities of 6G networks, which are still in the distant future, has long since begun. In addition to even faster access, shorter delays, and new applications, features such as human-centricity, data protection, and privacy are being addressed in particular in the discussions. These latter points make DIDs, SSI, and related concepts and architectures promising candidates for 6G adoption.The talk gives a brief introduction to DIDs and SSI and then discusses the benefits and drawbacks the integration of these technologies into 6G may have. Furthermore, the talk identifies different use cases and identifies the system components and functions of cellular networks affected by a 6G integration.

In this paper, we highlight a fundamental vulnerability associated with the widely adopted “Just Tap” push-based authentication in the face of a concurrency attack, and propose the method REPLICATE, a redesign to counter this vulnerability. In the concurrency attack, the attacker launches the login session at the same time the user initiates a session, and the user may be fooled, with high likelihood, into accepting the push notification which corresponds to the attacker's session, thinking it is their own. The attack stems from the fact that the login notification is not explicitly mapped to the login session running on the browser in the Just Tap approach. REPLICATE attempts to address this fundamental flaw by having the user approve the login attempt by replicating the information presented on the browser session over to the login notification, such as by moving a key in a particular direction, choosing a particular shape, etc. We report on the design and a systematic usability study of REPLICATE. Even without being aware of the vulnerability, in general, participants placed multiple variants of REPLICATE in competition to the Just Tap and fairly above PIN-based authentication.

In the last decade the use of artificial neural networks (ANNs) in many fields like image processing or speech recognition has become a common practice because of their effectiveness to solve complex tasks. However, in such a rush, very little attention has been paid to security aspects. In this work we explore the possibility to embed a watermark into the ANN parameters. We exploit model redundancy and adaptation capacity to lock a subset of its parameters to carry the watermark sequence. The watermark can be extracted in a simple way to claim copyright on models but can be very easily attacked with model fine-tuning. To tackle this culprit we devise a novel watermark aware training strategy. We aim at delving into the loss landscape to find an optimal configuration of the parameters such that we are robust to fine-tuning attacks towards the watermarked parameters. Our experimental results on classical ANN models trained on well-known MNIST and CIFAR-10 datasets show that the proposed approach makes the embedded watermark robust to fine-tuning and compression attacks.

It is not possible to speak about a complete security system without also taking into account the video surveillance system (CCTV). The reason is that CCTV systems offer the most spectacular results both in the security of goods and people and in the field of customer relations, marketing, traffic monitoring and the list can go on. With the development of the software industry the applicability of CCTV systems has greatly increased, largely due to image processing applications. The present paper, which is the continuation of an article already presented at an international conference, aims to shape the electrical characteristics of a common video surveillance system. The proposed method will be validated in two different programming environments.

in this paper, we describe a method for detecting abnormal pedestrians or cars by expressing the behavioral characteristics of pedestrians on a global surveillance map in a video security system using CCTV and patrol robots. This method converts a large amount of video surveillance data into a compressed map shape format to efficiently transmit and process data. By using deep learning auto-encoder and CNN algorithm, pedestrians belonging to the abnormal category can be detected in two steps. In the case of the first-stage abnormal candidate extraction, the normal detection rate was 87.7%, the abnormal detection rate was 88.3%, and in the second stage abnormal candidate filtering, the normal detection rate was 99.8% and the abnormal detection rate was 96.5%.

Privacy preserving scheme for face verification is a biometric system embedded with template protection to protect the data in ensuring data integrity. This paper proposes a new method called Histogram of Oriented Gradient Random Template Protection (HOGRTP). The proposed method utilizes Histogram of Oriented Gradient approach as a feature extraction technique and is combined with Random Template Protection method. The proposed method acts as a multi-factor authentication technique and adds a layer of data protection to avoid the compromising biometric issue because biometric is irreplaceable. The performance accuracy of HOGRTP is tested on the unconstrained face images using the benchmarked dataset, Labeled Face in the Wild (LFW). A promising result is obtained to prove that HOGRTP achieves a higher verification rate in percentage than the pure biometric scheme.

Generalizing automated medical image segmentation methods to new image domains is inherently difficult. We have previously developed a number of automated segmentation methods that perform at the level of human readers on images acquired under similar conditions to the original training data. We are interested in exploring techniques that will improve model generalization to new imaging domains. In this study we explore a method to limit the inherent bias of these models to intensity and textural information. Using a dataset of 100 T2-weighted MR images with fat-saturation, and 100 T2-weighted MR images without fat-saturation, we explore the use of neural style transfer to induce shape preference and improve model performance on the task of segmenting the kidneys in patients affected by polycystic kidney disease. We find that using neural style transfer images improves the average dice value by 0.2. In addition, visualizing individual network kernel responses highlights a drastic difference in the optimized networks. Biasing models to invoke shape preference is a promising approach to create methods that are more closely aligned with human perception.

Nowadays, we have witnessed the early progress on learning the association between voice and face automatically, which brings a new wave of studies to the computer vision community. However, most of the prior arts along this line (a) merely adopt local information to perform modality alignment and (b) ignore the diversity of learning difficulty across different subjects. In this paper, we propose a novel framework to jointly address the above-mentioned issues. Targeting at (a), we propose a two-level modality alignment loss where both global and local information are considered. Compared with the existing methods, we introduce a global loss into the modality alignment process. The global component of the loss is driven by the identity classification. Theoretically, we show that minimizing the loss could maximize the distance between embeddings across different identities while minimizing the distance between embeddings belonging to the same identity, in a global sense (instead of a mini-batch). Targeting at (b), we propose a dynamic reweighting scheme to better explore the hard but valuable identities while filtering out the unlearnable identities. Experiments show that the proposed method outperforms the previous methods in multiple settings, including voice-face matching, verification and retrieval.

Engineering drawings such as Piping and Instrumentation Diagrams contain a vast amount of text data which is essential to identify shapes, pipeline activities, tags, amongst others. These diagrams are often stored in undigitised format, such as paper copy, meaning the information contained within the diagrams is not readily accessible to inspect and use for further data analytics. In this paper, we make use of the benefits of recent deep learning advances by selecting models for both text detection and text recognition, and apply them to the digitisation of text from within real world complex engineering diagrams. Results show that 90% of text strings were detected including vertical text strings, however certain non text diagram elements were detected as text. Text strings were obtained by the text recognition method for 86% of detected text instances. The findings show that whilst the chosen Deep Learning methods were able to detect and recognise text which occurred in simple scenarios, more complex representations of text including those text strings located in close proximity to other drawing elements were highlighted as a remaining challenge.

Network port scans are a key first step to developing a true understanding of a network-facing attack surface. However in large-scale networks, the data resulting from such scans can be too numerous for Red Teams to process for manual and semiautomatic testing. Indiscriminate port scans can also compromise a Red Team seeking to quickly gain a foothold on a network. A large attack surface can even complicate Blue Team activities like threat hunting. In this paper we provide a cluster analysis methodology designed to group similar hosts to reduce security team workload and Red Team observability. We also measure the Internet-facing network attack surface of 13 organizations by clustering their hosts based on similarity. Through a case study we demonstrate how the output of our clustering technique provides new insight to both Red and Blue Teams, allowing them to quickly identify potential high-interest points on the attack surface.

This study extends previous advances in soft biometrics and describes to what extent soft biometrics can be used for facial profile recognition. The purpose of this research is to explore human recognition based on facial profiles in a comparative setting based on soft biometrics. Moreover, in this work, we describe and use a ranking system to determine the recognition rate. The Elo rating system is employed to rank subjects by using their face profiles in a comparative setting. The crucial features responsible for providing useful information describing facial profiles have been identified by using relative methods. Experiments based on a subset of the XM2VTSDB database demonstrate a 96% for recognition rate using 33 features over 50 subjects.

Clustering is a concept in data mining, which divides a data set into different classes or clusters according to a specific standard, making the similarity of data objects in the same cluster as large as possible. Clustering by fast search and find of density peaks (DPC) is a novel clustering algorithm based on density. It is simple and novel, only requiring fewer parameters to achieve better clustering effect, without the requirement for iterative solution. And it has expandability and can detect the clustering of any shape. However, DPC algorithm still has some defects, such as it employs the clear neighborhood relations to calculate local density, so it cannot identify the neighborhood membership of different values of points from the distance of points and It is impossible to accurately cluster the data of the multi-density peak. The fuzzy neighborhood density peak clustering algorithm is proposed for this shortcoming (F-DPC): novel local density is defined by the fuzzy neighborhood relationship. The fuzzy set theory can be used to make the fuzzy neighborhood function of local density more sensitive, so that the clustering for data set of various shapes and densities is more robust. Experiments show that the algorithm has high accuracy and robustness.

In recent year, there has been a growing need for intelligent systems that not only are able to provide reliable classifications but can also produce explanations for the decisions they make. The demand for increased explainability has led to the emergence of explainable artificial intelligence (XAI) as a specific research field. In this context, fuzzy logic systems represent a promising tool thanks to their inherently interpretable structure. The use of a rule-base and linguistic terms, in fact, have allowed researchers to create models that are able to produce explanations in natural language for each of the classifications they make. So far, however, designing systems that make use of interval type-2 (IT2) fuzzy logic and also give explanations for their outputs has been very challenging, partially due to the presence of the type-reduction step. In this paper, it will be shown how constrained interval type-2 (CIT2) fuzzy sets represent a valid alternative to conventional interval type-2 sets in order to address this issue. Through the analysis of two case studies from the medical domain, it is shown how explainable CIT2 classifiers are produced. These systems can explain which rules contributed to the creation of each of the endpoints of the output interval centroid, while showing (in these examples) the same level of accuracy as their IT2 counterpart.

For real-time applications of arbitrary style transformation, there is a trade-off between the quality of results and the running time of existing algorithms. Hence, it is required to maintain the equilibrium of the quality of generated artwork with the speed of execution. It's complicated for the present arbitrary style-transformation procedures to preserve the structure of content-image while blending with the design and pattern of style-image. This paper presents the implementation of a network using SANET models for generating impressive artworks. It is flexible in the fusion of new style characteristics while sustaining the semantic-structure of the content-image. The identity-loss function helps to minimize the overall loss and conserves the spatial-arrangement of content. The results demonstrate that this method is practically efficient, and therefore it can be employed for real-time fusion and transformation using arbitrary styles.

With the advancement of Artificial Intelligence technology to make smart devices, understanding how humans develop trust in virtual agents is emerging as a critical research field. Through our research, we report on a novel methodology to investigate user's trust in auditory assistance in a Virtual Reality (VR) based search task, under both high and low cognitive load and under varying levels of agent accuracy. We collected physiological sensor data such as electroencephalography (EEG), galvanic skin response (GSR), and heart-rate variability (HRV), subjective data through questionnaire such as System Trust Scale (STS), Subjective Mental Effort Questionnaire (SMEQ) and NASA-TLX. We also collected a behavioral measure of trust (congruency of users' head motion in response to valid/ invalid verbal advice from the agent). Our results indicate that our custom VR environment enables researchers to measure and understand human trust in virtual agents using the matrices, and both cognitive load and agent accuracy play an important role in trust formation. We discuss the implications of the research and directions for future work.

Neural Style Transfer based on convolutional neural networks has produced visually appealing results for image and video data in the recent years where e.g. the content of a photo and the style of a painting are merged to a novel piece of digital art. In practical engineering development, we utilize 3D objects as standard for optimizing digital shapes. Since these objects can be represented as binary 3D voxel representation, we propose to extend the Neural Style Transfer method to 3D geometries in analogy to 2D pixel representations. In a series of experiments, we first evaluate traditional Neural Style Transfer on 2D binary monochromatic images. We show that this method produces reasonable results on binary images lacking color information and even improve them by introducing a standardized Gram matrix based loss function for style. For an application of Neural Style Transfer on 3D voxel primitives, we trained several classifier networks demonstrating the importance of a meaningful convolutional network architecture. The standardization of the Gram matrix again strongly contributes to visually improved, less noisy results. We conclude that Neural Style Transfer extended by a standardization of the Gram matrix is a promising approach for generating novel 3D voxelized objects and expect future improvements with increasing graphics memory availability for finer object resolutions.

Trademarks are recognizable images and/or words used to distinguish various products or services. They become associated with the reputation, innovation, quality, and warranty of the products. Countries around the world have offices for industrial/intellectual property (IP) registration. A new trademark image in application for registration should be distinct from all the registered trademarks. Due to the volume of trademark registration applications and the size of the databases containing existing trademarks, it is impossible for humans to make all the comparisons visually. Therefore, technological tools are essential for this task. In this work we use a pre-trained, publicly available Convolutional Neural Network (CNN) VGG19 that was trained on the ImageNet database. We adapted the VGG19 for the trademark image retrieval (TIR) task by fine tuning the network using two different databases. The VGG19v was trained with a database organized with trademark images using visual similarities, and the VGG19c was trained using trademarks organized by using conceptual similarities. The database for the VGG19v was built using trademarks downloaded from the WEB, and organized by visual similarity according to experts from the IP office. The database for the VGG19c was built using trademark images from the United States Patent and Trademarks Office and organized according to the Vienna conceptual protocol. The TIR was assessed using the normalized average rank for a test set from the METU database that has 922,926 trademark images. We computed the normalized average ranks for VGG19v, VGG19c, and for a combination of both networks. Our method achieved significantly better results on the METU database than those published previously.