<p dir="ltr">VANET is formed by vehicles, road units, infrastructure components, and various con- nected objects.It aims mainly to ensure public safety and traffic control. New emerging applications include value-added and user-oriented services. While this technological ad- vancement promises ubiquitous deployment of the VANET, security and privacy challenges must be addressed. Thence, vehicle authentication is a vital process to detect malicious users and prevent them from harming legitimate communications. Hover, the authentication pro- cess uses sensitive information to check the vehicle’s identity. Sharing this information will harm vehicle privacy. In this thesis, we aim to deal with this issues:</p><ul><li>How can we ensure vehicle authentication and avoid sensitive and identity information leaks simultaneously?</li><li>When nodes are asked to provide identity proof, how can we ensure that the shared information is only used by an authorized entity?</li><li>Can we define an effective scheme to distinguish between legitimate and malicious network nodes?This dissertation aims to address the preservation of vehicle private information used within the authentication mechanism in VANET communications.The VANET characteristics are thoroughly presented and analyzed. Security require- ments and challenges are identified. Additionally, we review the proposed authentication techniques and the most well-known security attacks while focusing on the privacy preser- vation need and its challenges.To fulfill, the privacy preservation requirements, we proposed a new solution called Active Bundle AUthentication Solution based on SDN for Vehicular Networks (ABAUS). We intro- duce the Software Defined Networks (SDN) as an authentication infrastructure to guarantee the authenticity of each participant. Furthermore, we enhance the preservation of sensitive data by the use of an active data Bundle (ADB) as a self-protecting security mechanism. It ensures data protection throughout the whole data life cycle. ABAUS defines a dedicated registration protocol to verify and validate the different members of the network.</li></ul><p dir="ltr">first solution focused on legitimate vehicle identification and sensitive data pro- tection. A second scheme is designed to recognize and eliminate malicious users called BEhaviour-based REPutation scheme for privacy preservation in VANET using blockchain technology (BEREP). Dedicated public blockchains are used by a central trust authority to register vehicles and store their behavior evaluation and a trust scoring system allows nodes to evaluate the behavior of their communicators and detect malicious infiltrated users.</p><p dir="ltr">By enhancing sensitive data preservation during the authentication process and detect- ing malicious attempts, our proposed work helps to tackle serious challenges in VANET communications.</p>

Privacy preservation is important for machine learning and datamining, but measures designed to protect private information sometimes result in a trade off: reduced utility of the training samples. This thesis introduces a privacy preserving approach that can be applied to decision-tree learning, without concomitant loss of accuracy. It describes an approach to the preservation of privacy of collected data samples in cases when information of the sample database has been partially lost. This approach converts the original sample datasets into a group of unreal datasets, where an original sample cannot be reconstructed without the entire group of unreal datasets. This approach does not perform well for sample datasets with low frequency, or when there is low variance in the distribution of all samples. However, this problem can be solved through a modified implementation of the approach introduced later in this thesis, by using some extra storage.

Due to the rise of digitalization and the growing amount of data, ensuring the integrity and security of patient data has become increasingly vital within the healthcare industry, which has traditionally managed substantial quantities of sensitive patient and personal information. This bachelor's thesis focused on designing and implementing a secure data sharing infrastructure to protect the integrity and confidentiality of patient data. Synthetic data was used to enable access for researchers and students in regulated environments without compromising patient privacy. The project successfully achieved its goals by evaluating different privacy-preserving mechanisms and developing a machine learning-based application to demonstrate the functionality of the secure data sharing infrastructure. Despite some challenges, the chosen algorithms showed promising results in terms of privacy preservation and statistical similarity. Ultimately, the use of synthetic data can promote fair decision-making processes and contribute to secure data sharing practices in the healthcare industry. / Hälso- och sjukvårdsbranschen har länge varit en sektor som hanterar stora mängder känsliga patientdata och personuppgifter. Integriteten och säkerheten hos patientdata har blivit allt viktigare som en följd av ökad datavolym och digitalisering. Detta examensarbete fokuserade på att utforma och implementera en säker datadelning infrastruktur för att skydda integritet och sekretess för patientdata. Syntetisk data användes för att möjliggöra tillgång för forskare och studenter i reglerade miljöer utan att riskera patienters privatliv. Projektet lyckades genom att utvärdera olika integritetsbevarande mekanismer och skapa en maskininlärningsbaserad applikation för att visa den säkra datadelningsinfrastrukturens funktionalitet. Trots vissa utmaningar visade de valda algoritmerna lovande resultat i fråga om integritetsbevarande och statistisk likhet. Slutligen kan användningen av syntetiska data främja rättvisa beslutsprocesser och bidra till säkra datadelningspraxis inom hälso- och sjukvårdsbranschen.

Biometrics are widely used in identification systems, such as face, fingerprint, iris, etc. Polar code is the only code that can be strictly proved to achieve channel capacity, and it has been proved to be optimal for channel and source coding. In this degree project, our goal is to apply polar codes algorithms to biometric identification systems, and to design a biometric identification system with high identification accuracy, low system complexity, and good privacy preservation. This degree project has carried out specific and in-depth research in four aspects, following results are achieved: First, idea of polar codes is learnt, for example channel combination, channel splitting, successive cancellation decoding. The successive cancellation and successive cancellation list algorithm are also applied to encoding, which further realizes polar codes for source coding. Second, using autoencoder to process biometrics. Autoencoder is introduced to compress fingerprints into binary sequences of length 1024, it has 5 encoding layers and 12 decoding layers, achieved reconstruction error is 0.03. The distribution is close to Gaussian distribution, and compressed codes are quantized into binary sequences. Properties of sequences are similar with random sequences in terms of entropy, correlation, variance. Third, the identification system under Wyner-Ziv problem is studied with fingerprints. In enrollment phase, encoding algorithms are designed to compress biometrics, and in identification phase, decoding algorithms are designed to estimate the original sequence based on decoded results and noisy sequence. Maximum mutual information method is used to identify users. Results show that with smaller number of users, longer code length, smaller noise, then recognition error rate is lower. Fourth, human faces are used in the generated secret key system. After fully considering the trade off to achieve optimal results, in enrollment phase both public data and secure data are generated, in identification phase user’s index and secret key are estimated. A hierarchical structure is further studied. First, CNN is used to classify the age of faces, and then the generated secret key system is used for identification after narrowing the range. The system complexity is reduced by 80% and the identification accuracy is not reduced. / Biometriska kännetecken används i stor utsträckning i identifieringssystem, kännetecken såsom ansikte, fingeravtryck, iris, etc. Polär kod är den enda koden som strikt bevisats uppnå kanalkapacitet och den har visat sig vara optimal för kanal- och källkodning. Målet med detta examensarbete är att tillämpa polära kodalgoritmer på biometriska identifieringssystem, och att designa ett biometriskt identifieringssystem med hög identifieringsnoggrannhet, låg systemkomplexitet och bra integritetsskydd. Under examensarbetet har det genomförts specifik och djupgående forskning i fyra aspekter, följande resultat har uppnåtts: För det första introduceras idén om polära koder, till exempel kanalkombination, kanaluppdelning, successiv annulleringsavkodning. Algoritmerna för successiv annullering och successiv annulleringslista tillämpas även på kodning,vilket ytterligare realiserar polära koders användning för källkodning. För det andra används autoencoder för att bearbeta biometriska uppgifter. Autoencoder introduceras för att komprimera fingeravtryck till binära sekvenser med längden 1024, den har 5 kodningslager och 12 avkodningslager, det uppnådda rekonstruktionsfelet är 0,03. Fördelningen liknar en normaldistribution och komprimerade koder kvantiseras till binära sekvenser. Egenskaperna för sekvenserna liknar slumpmässiga sekvenser vad gäller entropi, korrelation, varians. För det tredje studeras identifieringssystemet under Wyner-Ziv-problemet med fingeravtryck. I inskrivningsfasen är kodningsalgoritmer utformade för att komprimera biometriska kännetecken, och i identifieringsfasen är avkodningsalgoritmer utformade för att estimera den ursprungliga sekvensen baserat på avkodade resultat och brusiga sekvenser. Maximal ömsesidig informationsmetod används för att identifiera användare. Resultaten visar att med ett mindre antal användare, längre kodlängd och mindre brus så är identifieringsfelfrekvensen lägre. För det fjärde används mänskliga ansikten i det genererade hemliga nyckelsystemet. Efter att noggrant ha övervägt kompromisser fullt ut för att uppnå det optimala resultatet genereras både offentlig data och säker data under registreringsfasen, i identifieringsfasen uppskattas användarens index och säkerhetsnyckel. En hierarkisk struktur studeras vidare. Först används CNN för att klassificera ålder baserat på ansikten och sedan används det genererade hemliga nyckelsystemet för identifiering efter att intervallet har begränsats. Systemkomplexiteten reduceras med 80% men identifieringsnoggrannheten reduceras inte.

The insider threat problem is extremely challenging to address, as it is committed by insiders who are trusted and authorized to access the information resources of the organization. The problem is further complicated by the multifaceted nature of insiders, as human beings have various motivations and fluctuating behaviours. Additionally, typical monitoring systems may violate the privacy of insiders. Consequently, there is a need to consider a comprehensive approach to mitigate insider threats. This research presents a novel insider threat prevention and prediction model, combining several approaches, techniques and tools from the fields of computer science and criminology. The model is a Privacy- Preserving, Context-Aware, Insider Threat Prevention and Prediction model (PPCAITPP). The model is predicated on the Fraud Diamond (a theory from Criminology) which assumes there must be four elements present in order for a criminal to commit maleficence. The basic elements are pressure (i.e. motive), opportunity, ability (i.e. capability) and rationalization. According to the Fraud Diamond, malicious employees need to have a motive, opportunity and the capability to commit fraud. Additionally, criminals tend to rationalize their malicious actions in order for them to ease their cognitive dissonance towards maleficence. In order to mitigate the insider threat comprehensively, there is a need to consider all the elements of the Fraud Diamond because insider threat crime is also related to elements of the Fraud Diamond similar to crimes committed within the physical landscape. The model intends to act within context, which implies that when the model offers predictions about threats, it also reacts to prevent the threat from becoming a future threat instantaneously. To collect information about insiders for the purposes of prediction, there is a need to collect current information, as the motives and behaviours of humans are transient. Context-aware systems are used in the model to collect current information about insiders related to motive and ability as well as to determine whether insiders exploit any opportunity to commit a crime (i.e. entrapment). Furthermore, they are used to neutralize any rationalizations the insider may have via neutralization mitigation, thus preventing the insider from committing a future crime. However, the model collects private information and involves entrapment that will be deemed unethical. A model that does not preserve the privacy of insiders may cause them to feel they are not trusted, which in turn may affect their productivity in the workplace negatively. Hence, this thesis argues that an insider prediction model must be privacy-preserving in order to prevent further cybercrime. The model is not intended to be punitive but rather a strategy to prevent current insiders from being tempted to commit a crime in future. The model involves four major components: context awareness, opportunity facilitation, neutralization mitigation and privacy preservation. The model implements a context analyser to collect information related to an insider who may be motivated to commit a crime and his or her ability to implement an attack plan. The context analyser only collects meta-data such as search behaviour, file access, logins, use of keystrokes and linguistic features, excluding the content to preserve the privacy of insiders. The model also employs keystroke and linguistic features based on typing patterns to collect information about any change in an insider’s emotional and stress levels. This is indirectly related to the motivation to commit a cybercrime. Research demonstrates that most of the insiders who have committed a crime have experienced a negative emotion/pressure resulting from dissatisfaction with employment measures such as terminations, transfers without their consent or denial of a wage increase. However, there may also be personal problems such as a divorce. The typing pattern analyser and other resource usage behaviours aid in identifying an insider who may be motivated to commit a cybercrime based on his or her stress levels and emotions as well as the change in resource usage behaviour. The model does not identify the motive itself, but rather identifies those individuals who may be motivated to commit a crime by reviewing their computer-based actions. The model also assesses the capability of insiders to commit a planned attack based on their usage of computer applications and measuring their sophistication in terms of the range of knowledge, depth of knowledge and skill as well as assessing the number of systems errors and warnings generated while using the applications. The model will facilitate an opportunity to commit a crime by using honeypots to determine whether a motivated and capable insider will exploit any opportunity in the organization involving a criminal act. Based on the insider’s reaction to the opportunity presented via a honeypot, the model will deploy an implementation strategy based on neutralization mitigation. Neutralization mitigation is the process of nullifying the rationalizations that the insider may have had for committing the crime. All information about insiders will be anonymized to remove any identifiers for the purpose of preserving the privacy of insiders. The model also intends to identify any new behaviour that may result during the course of implementation. This research contributes to existing scientific knowledge in the insider threat domain and can be used as a point of departure for future researchers in the area. Organizations could use the model as a framework to design and develop a comprehensive security solution for insider threat problems. The model concept can also be integrated into existing information security systems that address the insider threat problem / Information Science / D. Phil. (Information Systems)