Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

This book constitutes the refereed proceedings of the 16th International Conference on on Applied Cryptography and Network Security, ACNS 2018, held in Leuven, Belgium, in July 2018. The 36 revised full papers presented were carefully reviewed and selected from 173 submissions. The papers were organized in topical sections named: Cryptographic Protocols; Side Channel Attacks and Tamper Resistance; Digital Signatures; Privacy Preserving Computation; Multi-party Computation; Symmetric Key Primitives; Symmetric Key Primitives; Symmetric Key Cryptanalysis; Public Key Encryption; Authentication and Biometrics; Cloud and Peer-to-peer Security.

Computer science. --- Computer security. --- Data encryption (Computer science). --- Coding theory. --- Computers and civilization. --- Computers. --- Law and legislation. --- Computer Science. --- Systems and Data Security. --- Data Encryption. --- Coding and Information Theory. --- Computers and Society. --- Legal Aspects of Computing. --- Information Systems Applications (incl. Internet). --- Automatic computers --- Automatic data processors --- Computer hardware --- Computing machines (Computers) --- Electronic brains --- Electronic calculating-machines --- Electronic computers --- Hardware, Computer --- Computer systems --- Cybernetics --- Machine theory --- Calculators --- Cyberspace --- Civilization and computers --- Civilization --- Data compression (Telecommunication) --- Digital electronics --- Information theory --- Signal theory (Telecommunication) --- Computer programming --- Data encoding (Computer science) --- Encryption of data (Computer science) --- Computer security --- Cryptography --- Computer privacy --- Computer system security --- Computers --- Cyber security --- Cybersecurity --- Electronic digital computers --- Protection of computer systems --- Security of computer systems --- Data protection --- Security systems --- Hacking --- Informatics --- Science --- Protection --- Security measures --- Cryptology. --- Law and legislation --- Data encryption (Computer science) --- Information theory. --- Application software. --- Application computer programs --- Application computer software --- Applications software --- Apps (Computer software) --- Computer software --- Communication theory --- Communication

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Verschillende quantum algoritmen (bv. Shor's algoritme) breken veel van de huidige publieke sleutel cryptografie (bv. RSA, DH, ECDH, ECDSA). Een voorstel om aan post quantum publieke sleutel cryptografie te kunnen doen is CSIDH. Deze constructie kan echter aangevallen worden door het Kuperberg algoritme. Lang werd gedacht dat symmetrische cryptografie minder risico liep. Met de komst van Simons algoritme werden verschillende van deze symmetrische schema's gebroken. Een voorstel was om de bitsgewijze optelling te vervangen door een andere operatie zodat Simons algoritme niet meer werkte. Een voorstel van een andere operatie was de modulaire optelling. In het geval van de modulaire optelling kan in plaats van Simons algoritme het Kuperberg algoritme gebruikt worden. In deze thesis passen we het Kuperbergalgoritme toe en proberen we een optimale heuristiek te bekomen in een van de stappen van het algoritme.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Machine learning is a popular method to analyse data sets: through learning the structure of a data set it can make decisions and predictions for new unseen data. With the rising awareness of privacy, not all data can be shared freely with other parties. Since good models require a good training data set, there is a need for sharing data without disclosing it. A good solution is homomorphic encryption. With homomorphic encryption the data can be shared and homomorphic encryption allows for computation on this data without actually revealing it. In this work, we investigate how to train a least squares support vector machine on homomorphically encrypted data. We train a least squares support vector machine on three data sets by solving a linear system of equations using a linear, polynomial and RBF kernel. We test several methods to solve the system: computing the inverse of the matrix of the system, the gradient descent algorithm with a fixed step size, the gradient descent algorithm with an adaptive step size and the conjugate gradient algorithm. To make the methods compatible with homomorphic encryption, we avoid divisions by using the Newton-Raphson algorithm and avoid matrix inversion by using the Schulz scheme. Encrypting the data adds a small noise to it. Our experiments are performed on plaintext data, however, we simulate this noise by adding a small Gaussian noise to the data. The performed experiments indicate that using the Newton-Raphson algorithm to avoid divisions is not practical due to its high multiplicative depth. The gradient descent method with a fixed step size has the advantage of not having to compute the step size in each iteration, however, it is not clear how to choose an appropriate step size. A wrong step size can increase the multiplicative depth significantly as then a lot of iterations are needed to make the solution converge or a wrong step size can make the algorithm divergent. In the three data sets we used, the polynomial kernel performed significantly worse than the other two kernels, the multiplicative depth needed to obtain the classifier is the largest for all iterative methods. According to the performed experiments, approximating the inverse of the matrix with the Schulz scheme is the most promising method.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

A genome wide association study (GWAS) processes highly sensitive personal information and must therefore take appropriate measures to ensure the privacy of the genomic data. Currently, federated databases increase the significance of GWASs but also create a trusted third party (TTP). Even though this is a reasonable assumption, the TTP is a single point of failure that could be compromised by hackers. Previous research, aimed at replacing the TTP with a multi-party computation (MPC) system, had either the drawback of a semi-honest security model or only performed a small part of a full GWAS protocol. Therefore, this thesis aims at investigating the feasibility of an MPC system with active security on a full GWAS protocol. The first contribution of this thesis is the design, implementation and theoretical scalability investigation of a full GWAS protocol. The SCALE-MAMBA actively-secure-with-abort framework is used for the implementation. By efficiently precomputing data, rewriting equations and investigating several algorithms, the multiplicative complexity of the full GWAS protocol is improved. Additional care had to be taken to assure the numerical stability of the implementations. The second contribution of this thesis is an extensive evaluation of the implemented GWAS protocol. A key insight is the feasibility of protocols with a constant amount of multiplications. Contrarily, amongst other, principal component analysis (PCA) is found to be impractical due to its higher multiplicative complexity that scales with the number of investigated SNPs and the number of individuals. Lastly, specific suggestions for future work detail how to make PCA and general SCALE-MAMBA implementations more practical. This could alleviate the performance issues apparent in the evaluation of the implemented GWAS protocol.