With the ever-growing amount of image data on the web, much attention has been devoted to large scale image search. It is one of the most challenging problems in computer vision for several... Show moreWith the ever-growing amount of image data on the web, much attention has been devoted to large scale image search. It is one of the most challenging problems in computer vision for several reasons. First, it must address various appearance transformations such as changes in perspective, rotation and scale existing in the huge amount of image data. Second, it needs to minimize memory requirements and computational cost when generating image representations. Finally, it needs to construct an efficient index space and a suitable similarity measure to reduce the response time to the users. This thesis aims to provide robust image representations that are less sensitive to above mentioned appearance transformations and are suitable for large scale image retrieval. Although this thesis makes a substantial number of contributions to large scale image retrieval, we also presented additional challenges and future research based on the contributions in this thesis. Show less
Unambiguous sequence variant descriptions are important in reporting the outcome of clinical diagnostic DNA tests. The standard nomenclature of the Human Genome Variation Society (HGVS)... Show moreUnambiguous sequence variant descriptions are important in reporting the outcome of clinical diagnostic DNA tests. The standard nomenclature of the Human Genome Variation Society (HGVS) describes the observed variant sequence relative to a given reference sequence. We propose an efficient algorithm for the extraction of HGVS descriptions from two DNA sequences. Our algorithm is able to compute the HGVS~descriptions of complete chromosomes or other large DNA strings in a reasonable amount of computation time and its resulting descriptions are relatively small. Additional applications include updating of gene variant database contents and reference sequence liftovers. Next, we adapted our method for the extraction of descriptions for protein sequences in particular for describing frame shifted variants. We propose an addition to the HGVS nomenclature for accommodating the (complex) frame shifted variants that can be described with our method. Finally, we applied our method to generate descriptions for Short Tandem Repeats (STRs), a form of self-similarity. We propose an alternative repeat variant that can be added to the existing HGVS nomenclature. The final chapter takes an explorative approach to classification in large cohort studies. We provide a ``cross-sectional'' investigation on this data to see the relative power of the different groups. Show less
Collaborative innovation processes in unpredictable environments are a challenge for traditional management. But new demands in a global digital society push public and corporate leadership to... Show moreCollaborative innovation processes in unpredictable environments are a challenge for traditional management. But new demands in a global digital society push public and corporate leadership to collaborate ad hoc, without predictable goals and planned working rules. In this study, an actor-network approach (ANT) is combined with critical incident technique (CIT) to elaborate dynamic network principles for a new real-time foresight (RTF). Real-time foresight replaces traditional planning and strategic management in ad hoc multi-sector collaborations. Although ANT originates from science and technologies studies, it is here applied to a management problem due to ist ability to merge voluntaristic and evolutionary managerial components and micro- and macro perspectives. The investigation is placed in an exemplary management field of high dynamics: global disaster management. From process analysis and from comparison of three dynamic innovation networks that emerged around Indian coastal villages after Tsunami 2004, five dynamic network patterns are obtained which underly successful collaborative innovation processes. These dynamic structures build the agenda for a new real-time foresight, and for an instrument to evaluate in real-time the emergence of dynamic innovation networks (DINs). Show less
Many scientists are focussed on building models. We nearly process all information we perceive to a model. There are many techniques that enable computers to build models as well. The field... Show more Many scientists are focussed on building models. We nearly process all information we perceive to a model. There are many techniques that enable computers to build models as well. The field of research that develops such techniques is called Machine Learning. Many research is devoted to develop computer programs capable of building models (algorithms). Many of such algorithms exist, and these often consist of various options that subtly influence performance (parameters). Furthermore, there is mathematical proof that there exists no single algorithm that works well on every dataset. This complicates the task of selecting the right algorithm for a given task. The field of meta-learning aims to resolve these problems. The purpose is to determine what kind of algorithms work well on which datasets. In order to do so, we developed OpenML. This is an online database on which researches can share experimental results amongst each other, potentially scaling up the size of meta-learning studies. Having earlier experimental results freely accessible and reusable for others, it is no longer required to conduct time expensive experiments. Rather, researchers can answer such experimental questions by a simple database look-up. This thesis addresses how OpenML can be used to answer fundamental meta-learning questions. Show less
Preferences have always been present in many tasks in our daily lives. Buying the right car, choosing a suitable house or even deciding on the food to eat, are trivial examples of decisions that... Show morePreferences have always been present in many tasks in our daily lives. Buying the right car, choosing a suitable house or even deciding on the food to eat, are trivial examples of decisions that reveal information, explicitly or implicitly, about our preferences. The recent trend of collecting increasing amounts of data is also true for preference data. Extracting and modeling preferences can provide us with invaluable information about the choices of groups or individuals. In areas like e-commerce, which typically deal with decisions from thousands of users, the acquisition of preferences can be a difficult task. For these reasons, artificial intelligence (in particular, machine learning) methods have been increasingly important to the discovery and automatic learning of models about preferences. In this Ph.D. project, several approaches were analyzed and proposed to deal with the LR problem. Most of which has focused on pattern mining methods. Show less
This thesis proposes design methodologies and techniques in the context of embedded computing systems. In particular, it focuses on embedded streaming systems, i.e., systems that process a... Show more This thesis proposes design methodologies and techniques in the context of embedded computing systems. In particular, it focuses on embedded streaming systems, i.e., systems that process a continuous, possibly infinite stream of data from the environment. Typical examples of such systems are audio and video encoders and decoders. In order to achieve higher performance, nowadays embedded streaming systems are often implemented on execution platforms that contain multiple processors on a single chip. These execution platforms are called Multi-Processor Systems-on-Chip (MPSoCs). To exploit the parallelism available in MPSoCs, applications have to be decomposed in portions (also called tasks) that are inter-dependent, but can be executed in parallel. Each of these tasks is assigned to a certain processor of the system. This assignment of tasks to processors is called spatial scheduling of tasks, or task mapping. This thesis proposes techniques to optimize and adapt at run-time the mapping of tasks to processors, in order to achieve higher processor utilization, or energy efficiency, or to make the system fault tolerant. Show less
The main goal of this dissertation is to manage resource allocation in network engineering problems and to introduce efficient cooperative algorithms to obtain high performance, ensuring... Show moreThe main goal of this dissertation is to manage resource allocation in network engineering problems and to introduce efficient cooperative algorithms to obtain high performance, ensuring fairness and stability. Specifically, this dissertation introduces new approaches for resource allocation in Orthogonal Frequency Division Multiple Access (OFDMA) wireless networks and in smart power grids by casting the problems to the coalitional game framework and by providing a constructive iterative algorithm based on dynamic learning theory. Show less
With the rapid development of electronic commerce, logistics management has become more and more important in the procedure of supply chain management. The goal of logistics management is to... Show moreWith the rapid development of electronic commerce, logistics management has become more and more important in the procedure of supply chain management. The goal of logistics management is to satisfy the demands of customers while minimizing the use of resources of the whole process in logistics management from the point of origin to the point of consumption. The logistics management technology has been widely used in the field of engineering and contributes to reducing the total logistics cost. In this thesis, we focus on algorithms based on nature-inspired paradigms to solve dynamic logistics management problems. Results show that the proposed multiple ant system algorithm and the multi-objective cooperative particle swarm algorithm are able to produce good solutions for the vehicle routing problems and inventory routing problems not only in theory but also in practice. Show less
We address two problems in Software Engineering. The first problem is how to assess the severity of software defects? The second problem we address is that of studying software designs. Automated... Show moreWe address two problems in Software Engineering. The first problem is how to assess the severity of software defects? The second problem we address is that of studying software designs. Automated support for assessing the severity of software defects helps human developers to perform this task more efficiently and more accurately. We present (MAPDESO) for assessing the severity of software defects based on IEEE Standard Classification for Software Anomalies. The novelty of the approach lies in its use of uses ontologies and ontology-based reasoning which links defects to system level quality properties. One of the main reasons that makes studying of software designs challenging is the lack of their availability. We decided to collect software designs represented by UML models stored in image formats and use image processing techniques to convert them to models. We present the 'UML Repository' which contains UML diagrams (in image and XMI format) and design metrics. We conducted a series of empirical studies using the UML Repository. These empirical studies are a drop in the ocean empirical studies that can be conducted using the repository. Yet these studies show the versatility of useful studies that can be based on this novel repository of UML designs. Show less
Image analysis of objects in the microscope scale requires accuracy so that measurements can be used to differentiate between groups of objects that are being studied. This thesis deals... Show more Image analysis of objects in the microscope scale requires accuracy so that measurements can be used to differentiate between groups of objects that are being studied. This thesis deals with measurements in yeast biology that are obtained through microscope images. We study the algorithms and workflow of image analysis of yeast cells in order to understand and improve the measurement accuracy. The Saccharomyces cerevisiae cell is widely used as a model organism in the life sciences. It is essential to study the gene and protein behaviour within these cells, and consequently making it possible to find treatment and solutions for genetic and hereditary diseases. This is possible since many processes that occurs at the molecular level in this organism are similar to those in human cells. In the research group Imaging and Bioinformatics, we have developed a framework for analysis of yeast cells. This framework is intended to serve as a support for research in yeast biology. The framework is integrated in one application and presented via a GUI. The application integrates modules and algorithms including segmentation, measurement, analysis and visualization. Show less
Parallel programming has become essential for writing scalable programs on general hardware. Conceptually, every parallel program consists of workers, which implement primary units of sequential... Show moreParallel programming has become essential for writing scalable programs on general hardware. Conceptually, every parallel program consists of workers, which implement primary units of sequential computation, and protocols, which implement the rules of interaction that workers must abide by. As programmers have been writing sequential code for decades, programming workers poses no new fundamental challenges. What is new---and notoriously difficult---is programming of protocols. In this thesis, I study an approach to protocol programming where programmers implement their workers in an existing general-purpose language (GPL), while they implement their protocols in a complementary domain-specific language (DSL). DSLs for protocols enable programmers to express interaction among workers at a higher level of abstraction than the level of abstraction supported by today's GPLs, thereby addressing a number of protocol programming issues with today's GPLs. In particular, in this thesis, I develop a DSL for protocols based on a theory of formal automata and their languages. The specific automata that I consider, called constraint automata, have transition labels with a richer structure than alphabet symbols in classical automata theory. Exactly these richer transition labels make constraint automata suitable for modeling protocols. Show less