Datenkommunikation, Netzwerke

A framework for managing the security risks to customers on the mobile payment ecosystem in Tanzania

Master's Thesis from the year 2015 in the subject Computer Science - IT-Security, , course: Information Technology, language: English, abstract: The trend of mobile payment businesses in Tanzania is steadily growing with the potential to revolutionize cash dominant economy to cashless which remarks the growing of financialinclusion in the country. However, that development brings many issues regarding customer security in Tanzania due to insufficient knowledge of users, poor information delivery channels, lack of appropriate legal infrastructure and inappropriate security properties.This study intended to propose a framework for addressing the potential security risks to customers on the mobile payments in Tanzania. In fulfilling that objective, the study explored thepotential security risks to customers; the roles of the mobile payments stakeholders in securing customers; the legal gaps in customer security; the liability for securing customers and lastlyproposed a framework for managing the security risks to customers on the mobile payment.The research was a case study of the potential security risks to customers on the mobile payments in Tanzania and used exploratory and descriptive approaches to enable familiarizationof the potential security risks to customers on the mobile payments in Tanzania and better data analysis. The study employed purposive sampling techniqueto draw a sample and used a structured questionnaire tocollect data. Data were captured by using SPSS and analyzed with MS Excel. Both qualitative and quantitative research approaches were used for better interpretation and presentation of findings.By synthesizing the literature review and empirical findings, this thesis has proposed a framework for managing security risks to customers on mobile payments in Tanzania. The core elements of this framework include collaboration of stakeholders, risk management program, customer awareness program and legal and policy reforms. This study recommends legal and policy reforms to accommodate mobile payment customer security issues.

A Graph Theoretical Approach to Pruning Deep Neural Networks

Bachelor Thesis from the year 2024 in the subject Computer Sciences - Artificial Intelligence, grade: 100/100, Baden-Wuerttemberg Cooperative State University (DHBW) (Economics), course: Business Information Systems - Data Science, language: English, abstract: Imagine a world where deep learning models, despite their immense power, are no longer constrained by computational limitations. This vision fuels the innovative research presented herein, a journey into the realm of efficient deep neural networks through the lens of graph theory. This work introduces MLP-Rank, a groundbreaking method for network pruning that leverages the principles of weighted PageRank to identify and strategically remove redundant connections within multilayer perceptrons (MLPs). By representing the neural network architecture as a graph, the algorithm meticulously assigns importance scores to each connection, allowing for the targeted elimination of less crucial pathways, drastically reducing computational overhead without sacrificing accuracy. The core of this research delves into the algorithm's theoretical underpinnings, exploring its structural adaptations and modifications to the standard PageRank to optimize performance within neural network topologies. Rigorous experimentation across diverse datasets, including MNIST, Fashion-MNIST, and CIFAR-10, and various MLP architectures validates the efficacy of the MLP-Rank method, demonstrating significant improvements in inference speed and model compression. This exploration extends to a critical analysis of the theoretical assumptions against empirical results, bridging the gap between predicted and observed performance, and paving the way for future advancements in deep learning optimization. Discover how the synergy of graph theory and network pruning unlocks a new era of efficient, streamlined deep learning models, poised to revolutionize applications in resource-constrained environments, making AI more accessible and practical than ever before. This research is essential reading for anyone interested in Deep Neural Networks, Network Pruning, Graph Theory, Weighted PageRank, Inference Optimization, Sparsity, Accuracy, Speedup, MLP-Rank, and Model Compression, offering valuable insights into the future of efficient AI. The proposed methodology not only promises faster inference times but also contributes to the development of more sustainable and environmentally friendly AI solutions by reducing the energy footprint of deep learning models.

A Multi-Layered, Tamper-Resistant Security Appliance for Financial Terminals

A Study in Protocol Obfuscation Techniques and their Effectiveness

Master's Thesis from the year 2017 in the subject Computer Science - IT-Security, grade: 4, Oxford University, language: English, abstract: There are two main reasons for people to hide their network communication. Some want to maintain their privacy, while others need to access resources that might not be available to them due to restrictions applied by a censor.Monitoring of internet usage by the governments is not something new but now many countries have actively started blocking internet content for their citizens. Necessity is the mother of invention. Researchers and developers are working on ingenious ways of bypassing such monitoring and restrictions. In recent years, protocol obfuscation is becoming a burning topic especially because of evolution of some state level censors. Internet users want to freely access resources however censors are deployed by several countries around to world to restrict this access. As the protocol obfuscation techniques are gradually improving and so are the censors.Meanwhile, there are some internet users who do not live in countries that restrict their access to the internet resources in any way but these individuals are concerned about confidentiality of their internet use. Government agencies are frequently in the media over accusations of spying, on their own citizens as well as citizens of other countries, which is making many internet users alarmed about their anonymity. Both censors and circumvention tools are evolving continuously and gradually. The aim of this study was to analyse the protocol obfuscation techniques and systems available to ordinary internet users. The efficiency of these techniques and systems is assessed in the light of available research in respect of detecting obfuscated protocols. Once a research proves that an obfuscation technique can be successfully attacked it makes it easier to implement possible mitigations and the research also provides guidance to the developers of future protocol obfuscation systems so that they learn from the mistakes of others. It was noted that all the protocol obfuscation systems researched in this study are vulnerable to detection. In some cases, the detection does not even require the use of sophisticated resources.

Agiles Projektmanagement im Kontext Künstlicher Intelligenz

An Improved and Anonymous Three-factor Authentication Key Exchange Protocol for Wireless Sensor Networks

Master's Thesis from the year 2017 in the subject Computer Science - IT-Security, grade: A, , course: MSCS, language: English, abstract: The concepts of Internet of Things (IOT) show that everything in the global network is interconnected and accessible. In IOT environment Wireless Sensor Networks (WSNs) play a very important role because of its ubiquitous nature which use for wide range of applications like military surveillance, health care, environmental monitoring, agriculture etc. WSNs consisting of large numbers of sensor nodes which sensed the sensory information from the physical phenomena and forward the same to the cluster head or gateway node, sensor node having limited battery power and cannot be recharge after deployment. WSNs are resource constraints in terms of memory, energy, computational cost and communication speed. This thesis is focus to developed light weight user authentication and key agreement protocol to access the real time information from the IOT environment. Most recently Amin et al. find out various security limitations of the Farash et al. protocol and they extended the same protocol to improve its security weaknesses and claimed that the protocol is secure in all aspects. In this thesis we scrutinized the Amin et al. protocol which reveals that the protocol still having numerous security weaknesses such as user anonymity and user traceability attacks. In response to aforementioned security limitations we designed robust smartcard base threefactor user authentication and session key agreement scheme for WSNs environment. We analyzed the novel protocol formally and informally, formal security verification has done by using BAN-Logic which show that the scheme achieve mutual authentication and session key agreement among the participant entities. Furthermore, this protocol has also simulated in popular security tool ProVerif which simulation results show that the protocol is safe and withstand against all possible attacks including the above mentioned. The performance of this scheme is also compared with associated schemes the novel protocol can be applicable in real-life applications.

An Improved Modular Exponentiation To Resist Power Analysis Attacks

Master's Thesis from the year 2019 in the subject Computer Science - IT-Security, grade: A, , course: Computer Science and Engineering, language: English, abstract: Modular exponentiation has been the primary operation in most of the asymmetric cryptosystems. However, for large exponents which serves as the keys for encryption and decryption, conventional way to compute modular exponentiation is very time consuming. Hence, different binary methods based on bit wise "squaring-multiplication" operations is more widely used. However, these approaches are vulnerable to side-channel attacks, specially timing and power analysis attacks. Hence, these approaches are improved so that such attacks can be mitigated.Two processes have been designed based on "Multiply-Always" binary method which injects dummy multiplications so that the original multiplications can remain intact. The two processes are being invoked based on a random variable which generates either 0 or 1. With these randomized modular exponentiation, the proposed approach can resist simple and differential power analysis attacks to a large extent. The proposed approach has been implemented using conventional "squaring-multiplication" and "Montgomery-Ladder" methods. Further, the work has also been extended to CRT-RSA which is widely used for fast computation. The analysishas been done for different key sizes of 1024 bit, 1536 bit and 2048 bit RSA & CRT-RSA.

Analyse und effiziente Überwachung sicherheitskritischer Bitcoin-Adressen

Akademische Arbeit aus dem Jahr 2017 im Fachbereich Informatik - IT-Security, Note: 1,3, Hochschule Ruhr West, Sprache: Deutsch, Abstract: In dieser Arbeit werden sicherheitskritische Aspekte der Kryptowährung Bitcoin analysiert. Dabei werden ausschließlich Angriffe betrachtet, die den direkten Diebstahl von Bitcoins durch das Erlangen des privaten Schlüssels ermöglichen. Ziel ist dabei nicht die Übernahme eines bestimmten Kontos, sondern die Überwachung eines Raumes von Adressen. Der Raum besteht dabei aus liquiden Adressen oder Adressen, die aktiv benutzt werden. Die grundlegende Annahme dabei ist, dass genutzte Konten nicht gleichmäßig über den verfügbaren Schlüsselraum verteilt sind und es somit einen Unterschied zwischen der theoretischen und praktischen Sicherheit von Bitcoins gibt. Deswegen stellt das Paper Bitcoin spezifische Bruteforce Verfahren vor, die nach aktiven Bitcoin-Adressen in vordefinierten Teilen des Schlüsselraumes suchen. Ein weiteres entscheidendes Kriterium ist dabei, wie viele Adressen pro Sekunde durchsucht werden können. Zum einen, um eine möglichst große Menge von Adressen definieren und durchsuchen zu können und zum anderen, um möglichst viele Adressen, die regelmäßig genutzt werden, permanent überwachen zu können. Ein Fokus liegt dabei auf performante offline Bruteforce Angriffe mit Hilfe von modernen GPUs und effizienten Datenstrukturen.

Analysis of the BEM Code of Professional Guidelines

Akademische Arbeit aus dem Fachbereich Informatik - IT-Security, Asia Pacific University of Technology and Innovation, Sprache: Deutsch, Abstract: This paper aims to examine, analyze and apply the ethical theories and the BEM (Board of Engineers Malaysia) Code of Professional Conduct to solve ethical related issues in the practice of engineering.A standard exists in each engineering company that all the members should comply with, commit and apply. In general, no specific right or wrong option exists for each ethical circumstance, therefore engineers need to provide guidelines, frameworks, or reference in order to comprehend their duties. These principles are structured into a defined set of codes known as the Code of Conduct. In line with applicable laws, regulations, rules, and the highest standards of ethics and values, the Code of Conduct is intended by the complete person. It gives a framework for an ethical assessment of the worker and outlines the rights, duties, and obligations of members. For companies, shareholders, and other stakeholders, the topics covered in this code are of paramount importance. This Code shall enter into force as soon as possible. It shall be the obligation of all members to obey the terms of this code in spirit and in writing by a corporation or an organization practicing the Code of Conduct. To this day, many businesses use the Code of Conduct in Malaysia and the Malaysia engineering board more in context (BEM).

Anforderungen an Grid-basierte Intrusion Detection Systeme

Diplomarbeit aus dem Jahr 2008 im Fachbereich Informatik - IT-Security, Note: 1,3, Ludwig-Maximilians-Universität München (Informatik), Sprache: Deutsch, Abstract: Mit der größer werdenden Verbreitung von Grids muss gleichzeitig ein tragfähiges, organisationsübergreifendes Sicherheitsmanagement etabliert werden. Besonders auch im neuartigen und noch stetigem Wandel unterworfenen Umfeld von Grids sind hierfür Intrusion Detection Systeme von großer Bedeutung, wo sie idealerweise etwaige schon bestehende Sicherheitslösungen ergänzen. Mit dem weiteren Anstieg der Leistungsfähigkeit von Grids müssen diese geradezu als "magische" Anziehungspunkte aufgefasst werden, wobei bedingt durch die charakteristische große Komplexität vielfältige Szenarien missbräuchlicher Nutzung realitätsnah erscheinen. Eine ausgereifte Einbruchserkennung ist nicht zuletzt auch vor dem Hintergrund einer weiter zunehmenden industriellen Nutzung von hohem Wert für eine dauerhafte Überlebensfähigkeit von Grids.Nach einer grundlegenden Einführung in die Thematik wird ausführlich der Status Quo gridbasierter Ansätze zur Einbruchserkennung untersucht. Es zeigt sich, dass trotz einiger Bemühungen derzeit noch keine umfassende und praktikable Lösung vorgestellt wurde. Vielmehr konzentrieren sich die Ansätze jeweils nur auf einzelne Teilaspekte der gesamten Problematik im Grid, ohne vorher ein angemessenes Fundament und hier besonders die grundsätzlichen Anforderungen im Umfeld von Grids begründet zu haben.In der Folge widmet sich die Arbeit im Rahmen einer ausführlichen, strukturierten Analyse der Ermittlung wichtiger generischer Anforderungen an gridbasierte Intrusion Detection Systeme, wobei diese jeweils aus praxisnahen Anwendungsfällen abgeleitet werden. Auf diese Weise wird ein Katalog von 48 verschiedenen Anforderungen erarbeitet, die jeweils auch beschrieben und mit einer Erläuterung ihrer Relevanz versehen werden.Zusätzlich werden mögliche Angriffe auf Grids im Zuge einer grundlegenden Bedrohungsanalyse als Teildisziplin des Security Engineerings diskutiert, wobei vor allem eine klare Klassifizierung derselben auf hohem Abstraktionsgrad erfolgt.

Arquitetura da Mente Artificial. Como a IA Modela à Educação Contemporânea e Estabelece o Silêncio Semântico

Academic Paper from the year 2026 in the subject Computer Sciences - Artificial Intelligence, , language: Portuguese, abstract: O presente livro propõe uma imersão em diferentes perspectivas da inteligência artificial na educação, articulando desde os fundamentos técnicos até as implicações subjetivas da automação no ensino. Para oferecer uma visão abrangente e estruturada, o estudo irá se dedicar a quatro temas fundamentais que conectam a tecnologia à prática pedagógica. A historiografia da Inteligência Artificial (IA) transcende a mera cronologia de incrementos computacionais e sucessões de avanços técnicos, ela se manifesta como a materialização de uma teleologia humana voltada à mimetização e, em última instância, à transcendência das faculdades cognitivas do Homo sapiens. Ao analisar a trajetória descrita por MacCorduck (2004), em "Machines Who Think", observa-se que o desenvolvimento de sistemas autônomos estabelece um diálogo direto com as atuais transformações pedagógicas, onde a tecnologia deixa de ser um suporte infraestrutural para se tornar um agente mediador ativo no processo de ensino-aprendizagem. Essa transição não é meramente instrumental, mas profunda, pois enquanto as raízes lógicas da IA remontam aos autômatos da antiguidade, a contemporaneidade introduz o que Shimabukuro (2024) caracteriza como um ponto de ruptura ontológica.

Big Data, Machine, and Deep Learning

Bring your own Device. Einführungsoptimierung von BYOD unter Datenschutzaspekten im Unternehmen

Studienarbeit aus dem Jahr 2016 im Fachbereich Informatik - IT-Security, Note: 1,3, FOM Hochschule für Oekonomie & Management gemeinnützige GmbH, Stuttgart, Veranstaltung: IT-Architekturen & Sicherheitsmanagement, Sprache: Deutsch, Abstract: Das Ziel der wissenschaftlichen Ausarbeitung ist, Unternehmen, die personenbezogene Daten erheben, verarbeiten oder nutzen und vor einer Einführung des BYOD-Modells stehen, eine Möglichkeit aufzuzeigen, die Einführung insbesondere im Hinblick auf den Datenschutz optimal zu gestalten. Die Identifikation und Bewertung der durch den Einsatz von BYOD resultierenden Risiken und das Aufzeigen von Maßnahmen für eine Risikominimierung dienen zur Erreichung der Zielsetzung. Betrachtet werden insbesondere die Anforderungen des Datenschutzes gemäß Bundesdatenschutzgesetz. Die Umsetzung und praktische Evaluierung von erarbeiteten Lösungsansätzen in Form von Maßnahmen zur Einführungsoptimierung von BYOD erfolgt nicht in dieser Ausarbeitung.

Cloud Computing im Mittelstand. Vertrauensbasis und IT-Sicherheit als zentrale Erfolgsfaktoren

Bachelorarbeit aus dem Jahr 2014 im Fachbereich Informatik - IT-Security, Note: 2,3, Fachhochschule Bonn-Rhein-Sieg in Sankt Augustin, Sprache: Deutsch, Abstract: Ziel dieser Arbeit ist es herauszufinden, welche Maßnahmen für den erfolgreichen Einsatz des Cloud Computing in deutschen Unternehmen ergriffen werden müssen, um das Vertrauen der potentiellen Kunden zu gewinnen, die IT-Sicherheit zu gewährleisten und kontinuierlich zu verbessern. Neben den Herausforderungen und aktuellen rechtlichen Grundlagen des Cloud Computing beschäftigt sich diese wissenschaftliche Arbeit im Wesentlichen mit der IT-Sicherheit beim Einsatz von Cloud Computing in KMU. Dabei wird untersucht, wie eine bessere Vertrauensbasis zwischen CSP und Kunde hergestellt werden kann. Um einen besseren Einblick in das Thema zu erlangen, werden vorab die wichtigsten Grundlagen des Cloud Computing und der IT-Sicherheit in Kapitel 2 behandelt. Kapitel 3 gewährt einen Einblick in die statistische Entwicklung des Cloud Computing. In Kapitel 4 wird eine kritische Betrachtung des Cloud Paradigmas vorgenommen. Dafür werden die Herausforderungen von Cloud Computing für die Unternehmen untersucht und erläutert. Diese sind in Chancen, Risiken sowie die Mobilität unterteilt. Darüber hinaus wird ein Überblick über die gesetzlichen und vertraglichen Rahmenbedingungen gegeben. Kapitel 4 schließt mit einigen Praxisbeispielen der vergangenen Zeit ab. Darauf aufbauend wird im weiteren Verlauf dieser Arbeit die Planung und Organisation des Einsatzes sowie die Einführung von Cloud-Lösungen näher betrachtet. Hier wird vor allem auf die Vertrauensbildung der Unternehmen zu den CSP näher eingegangen. Ergänzend dazu wird in diesem Kapitel ein möglicher Leitfaden für die Überlegung, Planung und den Einsatz von Cloud Computing für KMU entwickelt, der den Unternehmen bei der Entscheidungsfindung, Anbieterauswahl und Vertragsgestaltung mit dem CSP helfen soll. Kapitel 6 schließt mit einer Aussage über die mögliche zukünftige Entwicklung des Vertrauens und der IT-Sicherheit des Cloud Computing ab.

Cloud computing. A security analysis

Literature Review from the year 2017 in the subject Computer Science - IT-Security, grade: A, Anglia Ruskin University, language: English, abstract: In this paper, we will have a closer look at cloud computing benefits and its security concerns. In addition, we will discuss ways to protect cloud infrastructures.Cloud computing is a new trend of emerging technologies that can help IT professionals to have on-demand, scalable, and broad access IT Services with the lowest possible cost. It is based on the sharing of computing resources among different tenants or consumers. This concept will help in decreasing the capex cost and convert them into opex costs. Furthermore, it helps cloud consumers to make use of the professional expertise of the Cloud service providers which will help him to build a solid infrastructure. However, it also introduces risks that were not known for traditional IT Professionals and this is because of the fact that physical resources are shared by more than Tennant and cloud consumers. Traditionally, information technology infrastructure is designed based on maximum expected capacity. Afterwards, upgrade based on growth in usage or products end of life that require more start-up and operation cost, especially when it requires a team of experts to design and operate this infrastructure. However, Sviokla (2009) argues that IT services can be offered on-demand over a network including infrastructures, platforms, software (Daylmai, 2015) and services such as DOS filtration and security for emails through a data centres that are scattered all over the world called Cloud (Armbrust et al., 2010).Mell and Grance (2011) state that cloud computing is enabling on-demand broad access over the network to computing resources such as servers, storage, applications and services that have the ability to be provisioned quickly with very low management overhead and provider interactions.

Codificando el Futuro

Consumer Loan Credit Risk Analyser Using Neural Networks

Master's Thesis from the year 2007 in the subject Computer Sciences - Artificial Intelligence, , language: English, abstract: Imagine a world where loan defaults are predicted with uncanny accuracy, safeguarding financial institutions and empowering responsible lending. This book delves into the cutting-edge application of artificial neural networks to revolutionize consumer loan credit risk assessment. Embark on a journey through the intricate landscape of machine learning, exploring the biological inspiration behind neural networks and their evolution into powerful predictive tools. This comprehensive work meticulously compares the performance of two prominent neural network architectures: feed-forward backpropagation and radial basis function networks, against traditional statistical methods, offering a balanced perspective on their strengths and limitations. Discover how these sophisticated algorithms are implemented using MATLAB's Neural Network Toolbox to construct a robust credit risk analysis system. Uncover the secrets of data preprocessing, network training, and performance evaluation, gaining invaluable insights into the practical aspects of building a real-world risk prediction model. This book provides a rigorous performance analysis, offering a statistical method for credit risk analysis and the experimental method using neural networks. Whether you're a seasoned data scientist, a finance professional, or an academic researcher, this book provides a holistic understanding of consumer loan credit risk, neural networks, feed-forward backpropagation, radial basis function networks, and machine learning techniques transforming the financial sector. Explore the future scope of these innovative technologies, uncovering the vast potential for applications in diverse domains beyond credit risk. Prepare to be captivated by the potential to reshape the future of finance through the power of intelligent algorithms and data-driven decision-making. This exploration provides an in-depth analysis of performance, offering justification for differences between experimental and statistical methods and highlighting the effectiveness of neural network approaches in credit risk prediction. Prepare to explore the synergy of statistical and neural network approaches and unlock unparalleled insights into the intricate world of financial risk management. The book elucidates the pathway toward accurate and ethical lending practices, all driven by the transformative force of artificial intelligence and a deep understanding of consumer loan dynamics.

Cost-Benefit Analysis of Deploying Generative AI Chatbots at Guaranty Trust Bank (GTBank), Nigeria

Research Paper (undergraduate) from the year 2025 in the subject Computer Sciences - Artificial Intelligence, grade: A, , language: English, abstract: This thesis explores the economic feasibility of deploying generative AI chatbots at Guaranty Trust Bank (GTBank) in Nigeria. Using a five-year cost-benefit analysis, it assesses financial returns, operational efficiency, and customer service improvements. The study identifies key cost categories-software, infrastructure, training, and maintenance-against measurable benefits like reduced labor costs and enhanced digital engagement. Results show a positive net present value and a strong benefit-cost ratio, with break-even projected within three to four years. The research concludes that AI chatbot deployment offers not only economic value but also strategic advantages in Nigeria's increasingly competitive digital banking landscape.

Cryptography and Steganography in Digital Images

Master's Thesis from the year 2009 in the subject Computer Science - IT-Security, grade: 19.61 / 20, , language: English, abstract: Unveil the invisible world of digital secrets! This groundbreaking work delves into the fascinating realm where cryptography and steganography intertwine to safeguard and conceal information within digital images. Explore the intricate dance between security and stealth as this book meticulously dissects the art of hiding data in plain sight, specifically within the ubiquitous JPEG image format. From the fundamental principles of cryptography, encompassing single-key and public-key systems, to the nuanced techniques of steganography and watermarking, this comprehensive study lays bare the methods used to protect sensitive data. Journey through the architecture of JPEG image compression, understanding its strengths and vulnerabilities, before diving deep into the world of JPEG-based steganographic methods. Discover the inner workings of algorithms like Jsteg, F3, F4, F5, and Mod4, with a spotlight on the F5 algorithm's permutative straddling and matrix encoding, alongside the unique features of the Mod4 method. Through rigorous experimental analysis, using Matlab implementations, the book unveils the embedding capacity and image quality trade-offs inherent in these methods. A thorough statistical evaluation exposes the strengths and weaknesses of each algorithm against steganalysis attacks. But the innovation doesn't stop there! The book pioneers a novel approach, introducing F5 Shrinkage Mapping (F5_MOS) to enhance the performance of the F5 algorithm, improving embedding capacity and bolstering resistance to detection. Whether you're a cybersecurity professional, a digital forensics expert, or an avid enthusiast of information security, this book provides invaluable insights into the cutting-edge techniques used to protect and conceal data in the digital age, offering a compelling exploration of cryptography, steganography, and the ever-evolving landscape of information hiding. This exploration into digital images, image compression, data hiding, and information security will arm you with the knowledge to navigate the complex world of steganography and steganalysis.

Cryptography and Steganography. A multilayer Data Security Approach

Document from the year 2021 in the subject Computer Science - IT-Security, grade: 2, , language: English, abstract: This book focuses on the implementation of Image steganography and modifying the existing technique to add more security to a normal steganography technique. The Book emphasizes various techniques used in steganography with methodology, algorithm, MAT LAB implementation and implementation in Java.There is data everywhere, in the form of images, text and audio/video files. Some data is very crucial or personal and needs to be kept confidential. For that reason, we have cryptography. Cryptography is the science of encrypting a message such that an unauthorized party cannot comprehend what that message means. Only those who possess the key can decrypt the message. There might be cases where just encrypting the information is not enough, since an encrypted message can raise suspicion. Steganography is the science of hiding a message completely inside another form of data so that the existence of the actual message is not revealed during communication. It is possible to hide the image in any of the digital formats whether it is images, videos or even audio files. Images are popular because they are really frequent on the internet and hence do not arouse suspicion. This book intends to give an overview cryptography with image steganography.

Cyberbedrohungen. Eine Analyse von Kriterien zur Beschreibung von Advanced Persistent Threats

In Abgrenzung zum einfachen Hacking werden hochspezialisierte und hochgradig anspruchsvolle, oft auf einen längeren Zeitraum angelegte Cyberangriffe als Advanced Persistent Threats - kurz APTs - bezeichnet. Mandiant sieht in einem APT eine "Threat Group" - also eine Gruppe von Personen, die zusammenarbeitet, um für sie interessante Netzwerke anzugreifen. Eine beweissichere Zuordnung eines Cyberangriffs zu einem konkreten Täter (beispielsweise einem staatlichen Akteur) ist nicht unmöglich, aber so komplex, dass sie in den wenigsten Fällen erfolgreich ist. Dennoch versucht man im Zuge dieser Zuordnung, der sogenannten Attribution, möglichst viele Merkmale eines konkreten Angriffs mit den bekannten Merkmalen eines Advanced Persistent Threats abzugleichen. Kann man einer APT immer mehr Angriffe zuordnen, wird das Bild dieser Gruppe Schritt für Schritt genauer. In einigen Fällen war es dadurch bereits möglich, Rückschlüsse auf konkrete Täter ziehen zu können.Das Ziel dieses Buchs ist es, aus Fachberichten und weiteren Quellen ein umfassendes Bild der Beschreibungskriterien für Advanced Persistent Threats zu erarbeiten und zu analysieren. Diese Kriterien werden in einem Fragenkatalog für APT-Angriffe zusammengestellt. Es erfolgt eine Bewertung der Aussagekraft der einzelnen Kriterien. Dazu werden zwei konkrete Beispiele besonders bekannter Cyberangriffsgruppen, APT 10 und APT 28, angeführt.Aus dem Inhalt: - Cyberangriffe; - Advanced Persistent Threat; - APT-Kriterien; - Cyberbedrohungen

Das "Denken" von kognitiven Systemen

Das Diskriminierungspotenzial algorithmischer Entscheidungen im Hinblick auf die Reproduktion von Rassismus

Studienarbeit aus dem Jahr 2024 im Fachbereich Informatik - Künstliche Intelligenz, Note: 1.0, Universität zu Köln, Sprache: Deutsch, Abstract: Diese Arbeit beschäftigt sich mit der Frage, inwiefern algorithmische Entscheidungsträger objektiv und vorurteilsfrei im Umgang mit marginalisierten Gruppen agieren. Dafür wird zunächst definiert, was genau unter der Künstlichen Intelligenz und was unter einem Algorithmus verstanden wird. Aufgrund des begrenzten Umfangs der Arbeit liegt der Fokus auf der Diskriminierungsform Rassismus, wobei auch dieser Begriff definiert wird. Nachdem der aktuelle Forschungsstand hinsichtlich des Zusammenhanges von Algorithmen und Rassismus beleuchtet wurde, werden verschiedene Maßnahmen und Strategien gegen die diskriminierende Wirkung von Algorithmen präsentiert, um mögliche Lösungsansätze zur Verbesserung der Objektivität und Vorurteilsfreiheit der KI zu diskutieren. Abschließend folgen Fazit und Ausblick. Durch die rasante Entwicklung der Künstlichen Intelligenz (KI) hat sich in den letzten Jahren ein weitreichenden Forschungsfeld entwickelt, welches nicht nur die Technologie, sondern auch unser tägliches Leben maßgeblich beeinflusst. Wie der Philosoph Rainer Mühlhoff bemerkt, sind KIs weitaus mehr als Technologien; sie sind "Konstitutionsbedingungen unseres Denkens, Fühlens und Handelns." (ebd. 2020) Von Navigationsgeräten im Auto, Sprachassistenten wie SIRI oder Alexa, die Gesichtserkennung bei Handys oder Sprachübersetzungen wie Google Translate - KIs sind allgegenwärtig. Spätestens seitdem im Jahr 2022 der Chatbot ChatGPT1 veröffentlicht wurde, werden Künstliche Intelligenzen gesamtgesellschaftlich kontrovers diskutiert. Einige sehen in einer KI die Möglichkeit, die Welt zu verbessern, andere wiederum warnen vor KIs und deren negativen Auswirkungen. So wird neben der Annahme, dass KIs erheblich an Zeit und Arbeit sparen können, auch hinsichtlich der Fairness argumentiert. Denn eine wissenschaftlich und gesellschaftlich weitverbreitete Annahme ist, dass KIs im Gegensatz zu Menschen dazu fähig seien, vorurteils- und verzerrungsfrei zu bewerten und Entscheidungen zu treffen. Beispielsweise wenn es um eine Vorsortierung von Bewerbung geht; dabei könne eine KI objektiv vorfiltern, wer potenziell für den Job geeignet sei, so die Annahme. Da 60 Prozent der deutschen Unternehmen überzeugt davon sind, dass die KI eine derwichtigsten Zukunftstechnologien überhaupt darstellt wird die Präsenz dieser unaufhaltsam wachsen, denn die KI wurde entwickelt, um zu bleiben.

Datensicherheit und Datenschutz im Cloud Computing. Fallstudie und kritische Analyse