Fakultät für Informatik und Mathematik
Refine
Year of publication
Document Type
- Doctoral Thesis (90)
- Article (9)
- Conference Proceeding (7)
- Master's Thesis (1)
- Other (1)
- Preprint (1)
- Report (1)
Has Fulltext
- yes (110)
Is part of the Bibliography
- no (110)
Keywords
- Computersicherheit (7)
- Semantic Web (4)
- Web Security (4)
- Browser Security (3)
- Datenschutz (3)
- Internet of Things (3)
- Kryptologie (3)
- Multimedia (3)
- Suchmaschine (3)
- Virtual Network Embedding (3)
Institute
The amount of audio, video and image data on the Web is immensely growing, which leads to data management problems based on the hidden character of Multimedia. Therefore the interlinking of semantic concepts and media data with the aim to bridge the gap between the Internet of documents and the Web of Data has become a common practice. However, the value of connecting media to its semantic meta data is limited due to lacking access methods and the absence of an adapted query language specialized for media assets and fragments. This thesis aims to extend the standard query language for the Semantic Web (SPARQL) with media specific concepts and functions. The main contributions of the work are an exhaustive survey on Multimedia query languages of the last 3 decades, the SPARQL extension specification itself and an approach for the efficient evaluation of the new query concepts. Additionally I elaborate and evaluate a meta data based media fragment similarity approach, which provides a basis for further language extensions.
Cryptography is the scientific study of techniques for securing information and communication against adversaries. It is about designing and analyzing encryption schemes and protocols that protect data from unauthorized reading. However, in our modern information-driven society with highly complex and interconnected information systems, encryption alone is no longer enough as it makes the data unintelligible, preventing any meaningful computation without decryption. On the one hand, data owners want to maintain control over their sensitive data. On the other hand, there is a high business incentive for collaborating with an untrusted external party.
Modern cryptography encompasses different techniques, such as secure multiparty computation, homomorphic encryption or order-preserving encryption, that enable cloud users to encrypt their data before outsourcing it to the cloud while still being able to process and search on the outsourced and encrypted data without decrypting it. In this thesis, we rely on these cryptographic techniques for computing on encrypted data to propose efficient multiparty protocols for order-preserving encryption, decision tree evaluation and kth-ranked element computation.
We start with Order-preserving encryption (OPE) which allows encrypting data, while still enabling efficient range queries on the encrypted data. However, OPE is symmetric limiting, the use case to one client and one server. Imagine a scenario where a Data Owner (DO) outsources encrypted data to the Cloud Service Provider (CSP) and a Data Analyst (DA) wants to execute private range queries on this data. Then either the DO must reveal its encryption key or the DA must reveal the private queries. We overcome this limitation by allowing the equivalent of a public-key OPE.
Decision trees are common and very popular classifiers because they are explainable. The problem of evaluating a private decision tree on private data consists of a server holding a private decision tree and a client holding a private attribute vector. The goal is to classify the client’s input using the server’s model such that the client learns only the result of the classification, and the server learns nothing. In a first approach, we represent the tree as an array and execute only d interactive comparisons (instead of 2 d as in existing solutions), where d denotes the depth of the tree. In a second approach, we delegate the complete tree evaluation to the server using somewhat or fully homomorphic encryption where the ciphertexts are encrypted under the client’s public key.
A generalization of a decision tree is a random forest that consists of many decision trees. A classification with a random forest evaluates each decision tree in the forest and outputs the classification label which occurs most often. Hence, the classification labels are ranked by their number of occurrences and the final result is the best ranked one. The best ranked element is a special case of the kth-ranked element. In this thesis, we consider the secure computation of the kth-ranked element in a distributed setting with applications in benchmarking and auctions. We propose different approaches for privately computing the kth-ranked element in a star network, using either garbled circuits or threshold homomorphic encryption.
We consider a number of enhancements to the standard neural network training paradigm. First, we show that carefully designed parameter update rules may replace the need for a loss function and its gradient. We introduce a parameter update rule that generalises the standard cross-entropy gradient, and allows directly controlling the relative effect of easy and hard examples on the training process. We show that the proposed update rule cannot be derived by using a loss function and yields better classification accuracy compared to training with the standard cross-entropy loss.
In addition, we study the effect of the loss function choice on the learnt representations. We introduce the Single Logit Classification (SLC) task: classifying whether a given class is the correct class for a given example, in a computationally efficient manner, based on the appropriate class logit alone. A natural principle is proposed, the Principle of Logit Separation (PoLS), as a guideline for choosing and designing loss functions suitable for the SLC task. We mathematically analyse the alignment of eleven existing and novel loss functions with this principle. Experiment results show that using loss functions that are aligned with this principle results in a representation in the logits layer in which each logit is more informative of its class correctness, leading to a considerably better SLC accuracy.
Further, we attempt to alleviate the dependency of standard neural network models on large amounts of quality labels. The task of weakly supervised one-shot detection is considered, in which at training time the model is trained without any localisation labels, and at test time it needs to identify and localise instances of unseen classes. We propose the attention similarity networks (ASN) for this task. ASN use a Siamese neural network to compute a similarity score between an exemplar and different locations in a target example. Then, an attention mechanism performs localisation by learning to attend to the correct locations. The ASN model outperforms the relevant baselines for weakly supervised one-shot detection tasks in the audio and computer vision domains.
Finally, we consider the problem of quantifying prediction confidence in the regression setting. We propose two novel algorithms for emitting calibrated prediction intervals for neural network regressors, at any given confidence level. The two algorithms require binning of the output space and training the neural network regressor as a classifier. Then, the calibration algorithms choose the intervals in the output space, making sure they contain the amount of posterior probability mass that results in the desired confidence level.
We have proposed a strategy for the creation of attributes based on hidden Markov models (HMM) characterizing the transaction from different points of view. This strategy makes it possible to integrate a broad spectrum of sequential information into the attributes of transactions. In fact, we model the authentic and fraudulent behavior of merchants and card holders according to two univariate characteristics: the date and the amount of transactions. In addition, attributes based on HMMs are created in a supervised manner, thereby reducing the need for expert knowledge for the creation of the fraud detection system. Ultimately, our HMM-based multi-perspective approach allows automated data pre-processing to model time correlations to complement and eventually replace transaction aggregation strategies to improve detection efficiency. Experiments carried out on a large set of credit card transaction data from the real world (46 million transactions carried out by Belgian card holders between March and May 2015) have shown that the strategy proposed for data preprocessing based on HMM can detect more fraudulent transactions when combined with the strategy of preprocessing reference data based on expert knowledge for the detection of credit card fraud.
Our subject of study is strong approximation of stochastic differential equations (SDEs) with respect to the supremum and the L_p error criteria, and we seek approximations that are strongly asymptotically optimal in specific classes of approximations. For the supremum error, we prove strong asymptotic optimality for specific tamed Euler schemes relating to certain adaptive and to equidistant time discretizations. For the L_p error, we prove strong asymptotic optimality for specific tamed Milstein schemes relating to certain adaptive and to equidistant time discretizations. To illustrate our findings, we numerically analyze the SDE associated with the Heston–3/2–model originating from mathematical finance.
In high-performance computing, one primary objective is to exploit the performance that the given target hardware can deliver to the fullest. Compilers that have the ability to automatically optimize programs for a specific target hardware can be highly useful in this context. Iterative (or search-based) compilation requires little or no prior knowledge and can adapt more easily to concrete programs and target hardware than static cost models and heuristics. Thereby, iterative compilation helps in situations in which static heuristics do not reflect the combination of input program and target hardware well. Moreover, iterative compilation may enable the derivation of more accurate cost models and heuristics for optimizing compilers. In this context, the polyhedron model is of help as it provides not only a mathematical representation of programs but, more importantly, a uniform representation of complex sequences of program transformations by schedule functions. The latter facilitates the systematic exploration of the set of legal transformations of a given program.
Early approaches to purely iterative schedule optimization in the polyhedron model do not limit their search to schedules that preserve program semantics and, thereby, suffer from the need to explore numbers of illegal schedules. More recent research ensures the legality of program transformations but presumes a sequential rather than a parallel execution of the transformed program. Other approaches do not perform a purely iterative optimization.
We propose an approach to iterative schedule optimization for parallelization and tiling in the polyhedron model. Our approach targets loop programs that profit from data locality optimization and coarse-grained loop parallelization. The schedule search space can be explored either randomly or by means of a genetic algorithm.
To determine a schedule's profitability, we rely primarily on measuring the transformed code's execution time. While benchmarking is accurate, it increases the time and resource consumption of program optimization tremendously and can even make it impractical. We address this limitation by proposing to learn surrogate models from schedules generated and evaluated in previous runs of the iterative optimization and to replace benchmarking by performance prediction to the extent possible.
Our evaluation on the PolyBench 4.1 benchmark set reveals that, in a given setting, iterative schedule optimization yields significantly higher speedups in the execution of the program to be optimized. Surrogate performance models learned from training data that was generated during previous iterative optimizations can reduce the benchmarking effort without strongly impairing the optimization result. A prerequisite for this approach is a sufficient similarity between the training programs and the program to be optimized.
Algebraic solving of polynomial systems and satisfiability of propositional logic formulas are not two completely separate research areas, as it may appear at first sight. In fact, many problems coming from cryptanalysis, such as algebraic fault attacks, can be rephrased as solving a set of Boolean polynomials or as deciding the satisfiability of a propositional logic formula. Thus one can analyze the security of cryptosystems by applying standard solving methods from computer algebra and SAT solving. This doctoral thesis is dedicated to studying solvers that are based on logic and algebra separately as well as integrating them into one such that the combined solvers become more powerful tools for cryptanalysis.
This disseration is divided into three parts. In this first part, we recall some theory and basic techniques for algebraic and logic solving. We focus mainly on DPLL-based SAT solving and techniques that are related to border bases and Gröbner bases. In particular, we describe in detail the Border Basis Algorithm and discuss its specialized version for Boolean polynomials called the Boolean Border Basis Algorithm.
In the second part of the thesis, we deal with connecting solvers based on algebra and logic. The ultimate goal is to combine the strength of different solvers into one. Namely, we fuse the XOR reasoning from algebraic solvers with the light, efficient design of SAT solvers. As a first step in this direction, we design various conversions from sets of clauses to sets of Boolean polynomials, and vice versa, such that solutions and models are preserved via the conversions. In particular, based on a block-building mechanism, we design a new blockwise algorithm for the CNF to ANF conversion which is geared towards producing fewer and lower degree polynomials. The above conversions allow usto integrate both solvers via a communication interface.
To reach an even tighter integration, we consider proof systems that combine resolution and polynomial calculus, i.e. the two most used proof systems in logic and algebraic solving. Based on such a proof system, which we call SRES, we introduce new types of solving algorithms that demostrate the synergy between Gröbner-like and DPLL-like solving. At the end of the second part of the dissertation, we provide some experiments based on a new benchmark which illustrate that the our new method based on DPLL has the potential to outperform CDCL SAT solvers.
In the third part of the thesis, we focus on practical attacks on various cryptograhic primitives. For instance, we apply SAT solvers in the case of algebraic fault attacks on the symmetric ciphers LED and derivatives of the block cipher AES. The main goal there is to derive so-called fault equations automatically from the hardware description of the cryptosystem and thus automatizate the attack. To give some extra power to a SAT solver that inverts the hash functions SHA-1 and SHA-2, we describe how to tweak the SAT solver using a programmatic interface such that the propagation of the solver and thus the attack itself is improved.
Internet browsers include Application Programming Interfaces (APIs) to support Web applications that require complex functionality, e.g., to let end users watch videos, make phone calls, and play video games. Meanwhile, many Web applications employ the browser APIs to rely on the user's hardware to execute intensive computation, access the Graphics Processing Unit (GPU), use persistent storage, and establish network connections.
However, providing access to the system's computational resources, i.e., processing, storage, and networking, through the browser creates an opportunity for attackers to abuse resources. Principally, the problem occurs when an attacker compromises a Web site and includes malicious code to abuse its visitor's computational resources. For example, an attacker can abuse the user's system networking capabilities to perform a Denial of Service (DoS) attack against third parties. What is more, computational resource abuse has not received widespread attention from the Web security community because most of the current specifications are focused on content and session properties such as isolation, confidentiality, and integrity.
Our primary goal is to study computational resource abuse and to advance the state of the art by providing a general attacker model, multiple case studies, a thorough analysis of available security mechanisms, and a new detection mechanism. To this end, we implemented and evaluated three scenarios where attackers use multiple browser APIs to abuse networking, local storage, and computation. Further, depending on the scenario, an attacker can use browsers to perform Denial of Service against third-party Web sites, create a network of browsers to store and distribute arbitrary data, or use browsers to establish anonymous connections similarly to The Onion Router (Tor). Our analysis also includes a real-life resource abuse case found in the wild, i.e., CryptoJacking, where thousands of Web sites forced their visitors to perform crypto-currency mining without their consent. In the general case, attacks presented in this thesis share the attacker model and two key characteristics: 1) the browser's end user remains oblivious to the attack, and 2) an attacker has to invest little resources in comparison to the resources he obtains.
In addition to the attack's analysis, we present how existing, and upcoming, security enforcement mechanisms from Web security can hinder an attacker and their drawbacks. Moreover, we propose a novel detection approach based on browser API usage patterns. Finally, we evaluate the accuracy of our detection model, after training it with the real-life crypto-mining scenario, through a large scale analysis of the most popular Web sites.
In various fields of image analysis, determining the precise geometry of occurrent edges, e.g. the contour of an object, is a crucial task. Especially the curvature of an edge is of great practical relevance. In this thesis, we develop different methods to detect a variety of edge features, among them the curvature.
We first examine the properties of the parabolic Radon transform and show that it can be used to detect the edge curvature, as the smoothness of the parabolic Radon transform changes when the parabola is tangential to an edge and also, when additionally the curvature of the parabola coincides with the edge curvature. By subsequently introducing a parabolic Fourier transform and establishing a precise relation between the smoothness of a certain class of functions and the decay of the Fourier transform, we show that the smoothness result for the parabolic Radon transform can be translated into a change of the decay rate of the parabolic Fourier transform.
Furthermore, we introduce an extension of the continuous shearlet transform which additionally utilizes shears of higher order. This extension, called the Taylorlet transform, allows for a detection of the position and orientation, as well as the curvature and other higher order geometric information of edges. We introduce novel vanishing moment conditions which enable a more robust detection of the geometric edge features and examine two different constructions for Taylorlets. Lastly, we translate the results of the Taylorlet transform in R^2 into R^3 and thereby allow for the analysis of the geometry of object surfaces.
The processing of personal information is omnipresent in our data-driven society enabling personalized services, which are regulated by privacy policies. Although privacy policies are strictly defined by the General Data Protection Regulation (GDPR), no systematic mechanism is in place to enforce them. Especially if data is merged from several sources into a data-set with different privacy policies associated, the management and compliance to all privacy requirements is challenging during the processing of the data-set. Privacy policies can vary hereby due to different policies for each source or personalization of privacy policies by individual users. Thus, the risk for negligent or malicious processing of personal data due to defiance of privacy policies exists.
To tackle this challenge, a privacy-preserving framework is proposed. Within this framework privacy policies are expressed in the proposed Layered Privacy Language (LPL) which allows to specify legal privacy policies and privacy-preserving de-identification methods. The policies are enforced by a Policy-based De-identification (PD) process. The PD process enables efficient compliance to various privacy policies simultaneously while applying pseudonymization, personal privacy anonymization and privacy models for de-identification of the data-set. Thus, the privacy requirements of each individual privacy policy are enforced filling the gap between legal privacy policies and their technical enforcement.