No description available.

Boolean satisfiability (SAT) is the problem of determining whether there exists an assignment of the Boolean variables to the truth values such that a given Boolean formula evaluates to true. SAT was the first example of an NP-complete problem. Only two decades ago SAT was mainly considered as of a theoretical interest. Nowadays, the picture is very different. SAT solving becomes mature and is a successful approach for tackling a large number of applications, ranging from artificial intelligence to industrial hardware design and verification. SAT solving consists of encodings and solvers. In order to benefit from the tremendous advances in the development of solvers, one must first encode the original problems into SAT instances. These encodings should not only be easily generated, but should also be efficiently processed by SAT solvers. Furthermore, an increasing number of practical applications in computer science can be expressed as constraint satisfaction problems (CSPs). However, encoding a CSP to SAT is currently regarded as more of an art than a science, and choosing an appropriate encoding is considered as important as choosing an algorithm. Moreover, it is much easier and more efficient to benefit from highly optimized state-of-the-art SAT solvers than to develop specialized tools from scratch. Hence, finding appropriate SAT encodings of CSPs is one of the most fascinating challenges for solving problems by SAT. This thesis studies SAT encodings of CSPs and aims at: 1) conducting a comprehensively profound study of SAT encodings of CSPs by separately investigating encodings of CSP domains and constraints; 2) proposing new SAT encodings of CSP domains; 3) proposing new SAT encoding of the at-most-one constraint, which is essential for encoding CSP variables; 4) introducing the redundant encoding and the hybrid encoding that aim to benefit from both two efficient and common SAT encodings (i.e., the sparse and order encodings) by using the channeling constraint (a term used in Constraint Programming) for SAT; and 5) revealing interesting guidelines on how to choose an appropriate SAT encoding in the way that one can exploit the availability of many efficient SAT solvers to solve CSPs efficiently and effectively. Experiments show that the proposed encodings and guidelines improve the state-of-the-art SAT encodings of CSPs.

NeuroEvolution besides deep learning is considered the most promising method to train and optimize neural networks. Neuroevolution uses genetic algorithms to train the controller of an agent performing various tasks. Traditionally, the controller of an agent will be encoded in a genome which will be directly translated into the neural network of the controller. All weights and the connections will be described by their elements in the genome of the agent. Direct Encoding – states if there is a single change in the genome it directly affects a change in the brain. Over time, different forms of encoding have been developed, such as Indirect and Developmental Encodings. This paper mainly concentrates on Developmental Encoding and how it could improve NeuroEvolution. The No-Free Lunch theorem states that there is no specific optimization method that would outperform any other. This does not mean that the genetic encodings could not outperform other methods on specific neuroevolutionary tasks. However, we do not know what tasks this might be. Thus here a range of different tasks is tested using different encodings. The hope is to find in which task domains developmental encodings perform best.

Boolean satisfiability (SAT) is the problem of determining whether there exists an assignment of the Boolean variables to the truth values such that a given Boolean formula evaluates to true. SAT was the first example of an NP-complete problem. Only two decades ago SAT was mainly considered as of a theoretical interest. Nowadays, the picture is very different. SAT solving becomes mature and is a successful approach for tackling a large number of applications, ranging from artificial intelligence to industrial hardware design and verification. SAT solving consists of encodings and solvers. In order to benefit from the tremendous advances in the development of solvers, one must first encode the original problems into SAT instances. These encodings should not only be easily generated, but should also be efficiently processed by SAT solvers. Furthermore, an increasing number of practical applications in computer science can be expressed as constraint satisfaction problems (CSPs). However, encoding a CSP to SAT is currently regarded as more of an art than a science, and choosing an appropriate encoding is considered as important as choosing an algorithm. Moreover, it is much easier and more efficient to benefit from highly optimized state-of-the-art SAT solvers than to develop specialized tools from scratch. Hence, finding appropriate SAT encodings of CSPs is one of the most fascinating challenges for solving problems by SAT. This thesis studies SAT encodings of CSPs and aims at: 1) conducting a comprehensively profound study of SAT encodings of CSPs by separately investigating encodings of CSP domains and constraints; 2) proposing new SAT encodings of CSP domains; 3) proposing new SAT encoding of the at-most-one constraint, which is essential for encoding CSP variables; 4) introducing the redundant encoding and the hybrid encoding that aim to benefit from both two efficient and common SAT encodings (i.e., the sparse and order encodings) by using the channeling constraint (a term used in Constraint Programming) for SAT; and 5) revealing interesting guidelines on how to choose an appropriate SAT encoding in the way that one can exploit the availability of many efficient SAT solvers to solve CSPs efficiently and effectively. Experiments show that the proposed encodings and guidelines improve the state-of-the-art SAT encodings of CSPs.

Context. Custom solutions to optical character recognition problems are able to reach higher recognition rates then a generic solution by their ability to exploiting the limitations in the problem domain. Such solutions can be generated with genetic algorithms. This thesis evaluates two different chromosome encodings on an optical character recognition problem with a limited problem domain. Objectives. The main objective for this study is to compare two different chromosome encodings used in a genetic algorithm generating neural networks for an optical character recognition problem to evaluate both the impact on the evolution of the network as well as the networks produced. Methods. A systematic literature review was conducted to find genetic chromosome encodings previously used on similar problem. One well documented chromosome encoding was found. We implemented the found hromosome ncoding called binary, as well as a modified version called weighted binary, which intended to reduce the risk of bad mutations. Both chromosome encodings were evaluated on an optical character recognition problem with a limited problem domain. The experiment was run with two different population sizes, ten and fifty. A baseline for what to consider a good solution on the problem was acquired by implementing a template matching classifier on the same dataset. Template matching was chosen since it is used in existing solutions on the same problem. Results. Both encodings were able to reach good results compared to the baseline. The weighted binary encoding was able to reduce the problem with bad mutations which occurred in the binary encoding. However it also had a negative impact on the ability of finding the best networks. The weighted binary encoding was more prone to enbreeding with a small population than the binary encoding. The best network generated using the binary encoding had a 99.65% recognition rate while the best network generated by the weighted binary encoding had a 99.55% recognition rate. Conclusions. We conclude that it is possible to generate many good solutions for an optical character problem with a limited problem domain. Even though it is possible to reduce the risk of bad mutations in a genetic lgorithm generating neural networks used for optical character recognition by designing the chromosome encoding, it may be more harmful than not doing it.

In contrast to a single clause a pseudo-Boolean (PB) constraint is much more expressive and hence it is easier to define problems with the help of PB constraints. But while PB constraints provide us with a high-level problem description, it has been shown that solving PB constraints can be done faster with the help of a SAT solver. To apply such a solver to a PB constraint we have to encode it with clauses into conjunctive normal form (CNF). While we can find a basic encoding into CNF which is equivalent to a given PB constraint, the solving time of a SAT solver significantly depends on different properties of an encoding, e.g. the number of clauses or if generalized arc consistency (GAC) is maintained during the search for a solution. There are various PB encodings that try to optimize or balance these properties. This thesis is about such encodings. For a better understanding of the research field an overview about the state-of-the art encodings is given. The focus of the overview is a simple but complete description of each encoding, such that any reader could use, implement and extent them in his own work. In addition two novel encodings are presented: The Sequential Weight Counter (SWC) encoding and the Binary Merger Encoding. While the SWC encoding provides a very simple structure – it is listed in four lines – empirical evaluation showed its practical usefulness in various applications. The Binary Merger encoding reduces the number of clauses a PB encoding needs while having the important GAC property. To the best of our knowledge currently no other encoding has a lower upper bound for the number of clauses produced by a PB encoding with this property. This is an important improvement of the state-of-the art, since both GAC and a low number of clauses are vital for an improved solving time of the SAT solver. The thesis also contributes to the development of new applications for PB constraint encodings. The programming library PBLib provides researchers with an open source implementation of almost all PB encodings – including the encodings for the special cases at-most-one and cardinality constraints. The PBLib is also the foundation of the presented weighted MaxSAT solver optimax, the PBO solver pbsolver and the WBO, PBO and weighted MaxSAT solver npSolver.

An ambitious long-term goal for neuroevolution, which studies how artificial evolutionary processes can be driven to produce brain-like structures, is to evolve neurocontrollers with a high density of neurons and connections that can adapt and learn from past experience. Yet while neuroevolution has produced successful results in a variety of domains, the scale of natural brains remains far beyond reach. In this dissertation two extensions to the recently introduced Hypercube-based NeuroEvolution of Augmenting Topologies (HyperNEAT) approach are presented that are a step towards more brain-like artificial neural networks (ANNs). First, HyperNEAT is extended to evolve plastic ANNs that can learn from past experience. This new approach, called adaptive HyperNEAT, allows not only patterns of weights across the connectivity of an ANN to be generated by a function of its geometry, but also patterns of arbitrary local learning rules. Second, evolvable-substrate HyperNEAT (ES-HyperNEAT) is introduced, which relieves the user from deciding where the hidden nodes should be placed in a geometry that is potentially infinitely dense. This approach not only can evolve the location of every neuron in the network, but also can represent regions of varying density, which means resolution can increase holistically over evolution. The combined approach, adaptive ES-HyperNEAT, unifies for the first time in neuroevolution the abilities to indirectly encode connectivity through geometry, generate patterns of heterogeneous plasticity, and simultaneously encode the density and placement of nodes in space. The dissertation culminates in a major application domain that takes a step towards the general goal of adaptive neurocontrollers for legged locomotion.

Graph Neural Networks and Transformers are very powerful frameworks for learning machine learning tasks. While they were evolved separately in diverse fields, current research has revealed some similarities and links between them. This work focuses on bridging the gap between GNNs and Transformers by offering a uniform framework that highlights their similarities and distinctions. We perform positional encodings and identify key properties that make the positional encodings node embeddings. We found that the properties of expressiveness, efficiency and interpretability were achieved in the process. We saw that it is possible to use positional encodings as node embeddings, which can be used for machine learning tasks such as node classification, graph classification, and link prediction. We discuss some challenges and provide future directions.

This paper will focus solely on the effectiveness of AV (antivirus) in detecting Metasploit payloads which have been encapsulated with different encapsulation modules. There seems to be a significant knowledge gap in the evaluation of commercial antivirus's software and their ability to detect malicious code and stop such code from being executed on IT systems. Therefore we would like to evaluate the capabilities of modern AV software with the use of penetration testing tools such as Metasploit. The research process is heavily reliant on a case study methodology as it can be argued that each payload generated reflects a case in itself. Firstly the payloads are generated and encapsulated through the self developed software, secondly they are uploaded to VirusTotal to be scanned with the use of their publicly available API, third the results are obtained from VirusTotal and stored locally. Lastly the results are filtered through with the software which in turn generates graphs of the results. These results will provide sufficient data in comparing encapsulation methods, payload detection rates, draw conclusions regarding which operating system may be most vulnerable as well as the overall state of modern AV software's capabilities in detecting malicious payloads. There are plenty of noteworthy conclusions to be drawn from the results, one of them being the most efficient encapsulation method powershell_base64 which had amongst the lowest detection rates in regards to the amounts of payloads it encoded, meaning that its encapsulation hid the malicious code from the AV at a higher degree than most the other encapsulation modules. The most noteworthy conclusion from the results gathered however is the encapsulation methods which obtained the absolute lowest detection rates, these were x86_nonalpha, x86_shikata_ga_nai, x86_xor_dynamic as well as payloads without any encoding at all, which had a few payloads reach among the lowest detection rates across the board (<20%).

There is a continuously growing demand for improved solutions that provide both efficient storage and efficient retrieval of big data for analytical purposes. This thesis researches the use of machine-learning together with feature engineering to recommend the most cost-effective compression algorithm and encoding combination for columns in a columnar database management system (DBMS). The framework consists of a cost function calculated using compression time, decompression time, and compression ratio. An XGBoost machine-learning model is trained on labels provided by the cost function to recommend the most cost-effective combination for columnar data within a column or vector-oriented DBMS. While the methods are applied on ClickHouse, one of the most popular open-source column-oriented DBMS on the market, the results are broadly applicable to column-oriented data which share data type and characteristics with IoT telemetry data. Using billions of available rows of numeric real business data obtained at Axis Communications in Lund, Sweden, a set of features are engineered to accurately describe the characteristics of a given column. The proposed framework allows for weighting the business interests (compression time, decompression time, and compression ratio) to determine the individually optimal cost-effective solution. The model reaches an accuracy of 99% on the test dataset and an accuracy of 90.1% on unseen data by leveraging data features that are predictive of compression algorithms and encodings performances. Following ClickHouse strategies and the most suitable practices in the field, combinations of general-purpose compression algorithms and data encodings are analysed that together yield the best results in efficiently compressing the data of certain columns. Applying the unweighted recommended combinations on all columns, the framework’s performance impact was measured to increase the average compression speed by 95.46%. Reducing the time to compress the columns from 31.17 seconds to compress the data to 13.17 seconds. Additionally, the decompression speed was increased by 59.87%, reducing the time to decompress the columns from 2.63 seconds to 2.02 seconds, at the cost of decreasing the compression ratio by 66.05%. Increasing the storage requirements by 94.9 MB. In column and vector databases, chunks of data belonging to a certain column are often stored together on a disk. Therefore, choosing the right compression algorithm can lower the storage requirements and boost database throughput.