Publication Server of Constructor University Library
Not a member yet
858 research outputs found
Sort by
Representing logics and logictranslations
No full textLogic is the study of formal languages for propositions and truth. Logics are used both as a foundation of mathematics and as specification languages in mathematics and computer science. Since logic is intricately intertwined with the nature of mathematics, the question how to represent logics in our minds is a constant challenge to our understanding. And only when it is understood can we begin to answer the corresponding question about logic translations. At the same time logics are used to a large extent in computer science to reason about both mathematics and software systems. This brings up the question how logics and their translations can be represented in a computer system. Throughout the 20th century several answers have already been provided. Most of them can be grouped into two kinds, which can be denoted by set/model theory and type/proof theory. We base our investigation on the desire to reconcile these two views. Our focus is on extending the existing notions of logic and logic translation in a way that retains their nature and the accumulated knowledge about them while leading them into a new direction. Our goal is to enrich both research fields by applying them more cogently to and making them more accessible and understandable to one another. While this answers the question how to represent logics and logic translations in our minds, it is not adequate for the specific constraints and use cases of software systems. Therefore, in a second investigation, we explore how to make our representations more concrete and robust enough to permit a mechanized treatment on a large scale. Taking these two results together, we obtain a triangle of different mathematical communities, research objectives, and philosophies consisting of set/model theory, proof/type theory, and mathematical knowledge management. Our main contribution is to integrate its corners into a coherent framework centered around logic that capitalizes on their comparative advantages
A Comparative Benchmark of Large Objects in Relational Databases
No full textOriginally Binary Large Objects (BLOBs) in databases were conceived as a means to capture any large data (whatever large meant at the time of writing) which, for whatever reason, cannot or should not be modeled relationally. Today we find images, movies, XML, formatted documents, and many more data types stored in database BLOBs. A particular challenge obviously is moving such large units of data as fast as possible, hence performance benchmarks are of interest.
However, while extensive evaluations have been undertaken for a variety of SQL workloads, BLOBs have not been the target of thorough benchmarking up to now. TPC and SPC-2 standards do not address BLOB benchmarking either.
We present a comparative BLOB benchmark of the leading commercial and open-source systems available under Unix/Linux. Commercial DBMSs are anonymised, open-source DBMSs benchmarked are PostgreSQL and MySQL.
Measurements show large differences between the systems under test, depending on various parameters. A surprising result is that overall the open-source DBMSs in most situations outperform commercial systems if configured wisely
Production of two major histocompatibility complex class I allotypes with different tapasin dependence and in vitro peptide binding studies
No full textMajor histocompatibility complex (MHC) class I complexes present antigenic peptides from intracellular proteins on the cell surface to T lymphocytes. The antigenic peptides are loaded onto MHC class I in the endoplasmic reticulum. Most MHC class I allotypes require the assistance of the peptide loading complex (PLC) for optimal peptide loading. During their association with the PLC, the average affinity of the peptide population increases until the MHC class I/peptide complexes travel to the cell surface. This so-called peptide editing depends on the presence of tapasin, a member of the PLC. The precise role of tapasin, however, is still controversial and its molecular mechanism remains unknown. Two closely related MHC class I allotypes show a strikingly different dependence on tapasin. In the absence of tapasin, HLA-B*4402 is strongly impaired in efficient peptide loading and cell surface expression, in contrast to HLA-B*4405. It has been suggested that the two allotypes have different molecular properties in the absence of peptide but no comparative in vitro study of them exists so far. In this work, single chain HLA-B*4402 and HLA-B*4405 are produced in different expression systems in order to compare the peptide binding behavior of the two allotypes in vitro. In bacteria, the proteins were incorporated into inclusion bodies as expected. Heterologous protein expression in P. pastoris, which has not been reported for MHC class I so far, led to hyperglycosylation and precipitation. The hyperglycosylation phenotype could be remedied by modification of the yeast glycosylation pathway, but both the bacterial and the yeast proteins could not be efficiently refolded. Single chain HLA-B*4405 produced in insect cells was able to bind peptides and first peptide binding kinetics could be obtained. Yet in none of the expression systems native versions of both proteins were produced in sufficient amounts, indicating that a more sophisticated expression system is required
Realization and characterization of microcrystalline silicon thin-film transistors
No full textThin-film transistors (TFTs) are the key element in the low cost large area electronics such as flat panel displays and sensor arrays. TFTs based on amorphous and polycrystalline silicon are the current industrial standard. However, amorphous silicon transistors are limited by the low charge carrier mobility and device stability, whereas polycrystalline silicon transistors are still relatively expensive due to the needs for additional annealing and/or crystallization steps. Microcrystalline silicon transistors have recently emerged to combine the distinctive advantages of both technologies with the exclusion of their disadvantages. In this thesis, transistors with distinctively high charge carrier mobilities exceeding 50 cm2/Vs were realized, which are more than 50 times higher than amorphous silicon transistors. Several aspects were investigated following the realization of transistors with high charge carrier mobility. These include the correlation of the material properties with the device properties, the influence of contact effects and post-fabrication thermal annealing on the transistor performances. In addition to these, inverter circuits based on microcrystalline silicon transistors were realized and the influence of the transistor parameters on the performance of the inverters were investigated
Development of Two Novel Fluorescent Enzyme Assays - Nano-TRF and Supramolecular Tandem Assays
No full textThis doctoral thesis describes two novel fluorescent methods for assaying enzymatic activity, which are referred to as nanosecond time-resolved fluorescence (Nano-TRF) assays and supramolecular tandem assays. The first method, developed in collaboration with Fa. Hoffmann-La Roche, introduces a novel fluorescent probe (Dbo) for enzyme assays, which combines several desirable properties. First, Dbo has an exceedingly long fluorescence lifetime, which allows the use of Nano-TRF detection to increase the robustness of an assay by suppressing background fluorescence, for example from library compounds in high-throughput screening (HTS). Second, Dbo is efficiently quenched by tryptophan and tyrosine, which allows single-label assays. And third, Dbo has a very small size and hydrophilicity compared to common aromatic hydrophobic fluorescent probes. It has been demonstrated that the combination of these properties affords a minimally invasive, yet very powerful approach to determine the activity of proteases, tyrosine kinases and phosphatases. The second method introduces the use of water-soluble macrocycles and fluorescent dyes for enzyme assays, which presents an economic, convenient, and general assay principle. The assay is based on the competition of dye versus substrate and product in the reversible formation of a complex with the macrocycle. The enzyme thus converts a weak competitor (substrate) into a strong competitor (product) or vice versa, which leads to a different fraction of fluorescent dye bound to the macrocycle. Depending on whether the macrocycle/dye-complex is more or less fluorescent than the free dye, an increase or a decrease in fluorescence results, which signals the enzymatic activity. The method was applied to amino acid decarboxylases and arginase, and the possibility to derive enzyme kinetic parameters and inhibition constants has been demonstrated. Furthermore, the assays were conceptualized by supporting simulation
A Mathematical Semantic Web
No full textThe body of mathematical knowledge is rapidly increasing and constantly changing. Zentralblatt MATH, an abstracting and reviewing service in the field of mathematics, maintains a database of more than 1.6 million mathematical documents and reports an annual growth by 80,000 articles. Similarly, the open internet archive for for electronic preprints of scientific papers, arXiv.org, contains close to half a million documents. These figures suggest that neither a mathematician's memory nor the time she devotes to studying new publications can possibly suffice to cover a significant fraction of the accumulated wealth of mathematical knowledge. In addition to the exponential increase in the amount of information, one also observes an increase in the complexity of mathematical content with more interdependencies between different areas within and beyond mathematics.
We propose a Mathematical Semantic Web to support mathematicians in efficiently managing and retrieving mathematical knowledge using computer systems and the internet. As with the World Wide Web, the Mathematical Semantic Web will allow authors of mathematics to publish their documents online which cumulate to a gigantic, decentralized and dynamic mathematical knowledge base. Authors semantically annotate their work in a special logical formalism, namely Description Logics, to allow computers to understand the actual knowledge contained therein. Based on these annotations, computer agents reason about the mathematical knowledge and provide novel services on the Mathematical Semantic Web, giving mathematicians efficient access to vast repositories of mathematics.
This thesis first analyzes the utility of Description Logics for formalizing mathematical knowledge. This analysis concludes with the proposal of a Mathematical Semantic Web which is introduced in great detail subsequently. We elaborate on the architecture and individual building blocks before describing the authoring process for the Mathematical Semantic Web. To motivate the value of our proposal, some services operating on the Mathematical Semantic Web are specified. In an effort to improve knowledge retrieval even further, we introduce the combination of domain and structural semantics for reasoning processes, thereby effectively leveraging additional knowledge.
Finally, we provide a list of Best Practices which aim at simplifying the knowledge modeling process and improving the quality of the resulting knowledge base. The Best Practices have been distilled from the experience gained during our case studies
Input-Output OOMs
No full textInput-output OOMs (IO-OOMs) are an extension of the basic OOM theory to deal with controlled stochastic processes. They were first presented in a comprehensive OOM tutorial text, where an early version of the OOM learning algorithm was partially transferred. Since then, OOM research has focussed on basic OOMs, in particular on developing statistically efficient learning algorithms, and research on IO-OOMs has been dormant. In this report it is shown how the main theorems for OOMs can be transferred to the case of IO-OOMs and how one of the current OOM learning algorithms can be adapted to learning IO-OOMs, yielding the first complete IO-OOM learning algorithm
Bilateral and multilateral cooperation in international trade : a political-economy perspective
No full textThere is widespread consensus in the literature on trade cooperation that bilateralism is a deficient form of cooperation compared to multilateralism. This consensus is regularly echoed by warnings from policymakers and the media when negotiations in the multilateral GATT/WTO stagnate. This volume puts these warnings about bilateral cooperation into a historical perspective. Its main goal is to shed light on the role of bilateralism and multilateralism in trade cooperation from 1860 to the present. On the basis of a theory-driven empirical analysis, the volume shows that bilateralism plays an important role in promoting liberalization. Bilateralism renders liberal trade politically feasible on the domestic level because it allows political actors to manage the domestic political costs and benefits that arise from economic actors who lose and win by commercial collaboration. However, bilateralism is not without problems. Concerns about domestic distribution create enforcement problems that might result in a halt or even a reversal of liberalization. These enforcement problems can be eliminated through multilateral cooperation. The downside of multilateralism is, however, that it is impossible to control the domestic political effects of trade cooperation. For these reasons, there is no single best form of cooperation and political actors face a dilemma in institutional choice. In the historical perspective, it can be further shown that bilateralism is more frequently chosen for trade liberalization than multilateralism. This finding indicates that bilateral cooperation is less of a threat to liberal trade than the conventional perspective suggests. The volume makes some suggestions about how an appropriately-equipped WTO could render bilateralism a real alternative to multilateralism
Membrane-protein matrix for technical application
No full textDuring this thesis three different projects have been combined to gain insight into the technical application of Membrane-Protein Matrices. First, a miniaturized, automated "patch-clamp-on-a-chip" system has been adapted for measurements on planar lipid bilayers. The functional setup shows the expected advantages of low stray capacitances with determined 10pF, and complete integration into the data processing environment. Using giant unilammellar vesicles for automated giga-seal formation over a micrometer-sized aperture in a glass chip, proved to be a promising technique for further automation and parallelization to facilitate high-throughput screening of pharmaceutical substances. Second, a calcein efflux assay for liposomes provides an analytical method to investigate the functional incorporation of pore forming proteins into lipid bilayers. It allows quick testing the quality of porin batches prior to time consuming single molecule measurements on planar lipid membranes. Donnan potential experiments on liposomes revealed the competition of two processes in the formation of a zeta potential. It is determined by the concentration gradient of the encapsulated polyelectrolyte and by the number as well as by the charge of the incorporated porins. The Donnan potential allows to couple otherwise uncharged liposomes to an external electric field. Furthermore the Donnan potential gives us another tool to investigate the permeability of lipid membranes. By the characterization of the Donnan potential we can study the diffusion of ions across lipid membranes through channel proteins as well as the permeability of porins for macromolecules. Third, during the industry project "Development of a biological fuel-cell", a battery test bed has been set up and operated. Measurements were done with a zinc/air battery, incorporating PowerZyme's proprietary Active Transport Membrane. Several graphical user interfaces for data acquisition and data analyses have been created
Performance analysis of multihop ad hoc and hybrid wireless networks
No full textMultihop mobile wireless networks is a subject of intense research interest for next generation wireless systems. As we move towards fourth generation (4G) networks, high data-rates of the order of 100 Mbps can be achieved only over short distances. A multihop network has smaller transmission distance as compared to an equivalent single-hop network. This enables the radio terminals that are spatially well-separated from each other to reuse the same resources when the resulting interferences are not too severe. In this dissertation, the multihop network design is studied for two different kinds of systems: multihop ad hoc network and multihop hybrid cellular network. In the first part of the thesis work, a multihop ad hoc network is analyzed in detail and the system capacity is computed. However, the calculation of optimal resource allocation that would maximize the system capacity is found to be an NP-hard problem. Considering this complexity, a random data hopping technique has been proposed over a time-slot (TS) partitioned system in order to improve the system capacity of the network. It has been found that the granularity of the TS(s) can be increased by reducing the duration of each TS. Significantly, applying the random data hoping technique over the TS partitioned system results in an increase in the system capacity without any additional increase in the computational complexity. In the second part of the thesis work, a multihop hybrid cellular network is studied. It is known that optimal resource allocation in a multihop cellular network is again a NP-hard problem. A novel cluster-based design is proposed in this work for a 2-hop cellular network, whereby, two pairs communicate simultaneously in every hexagonal cell, at any time instant. Such a cluster-based 2-hop design provides an increase in the spatial resource reuse and hence results in a much higher system capacity as compared to three standard benchmark algorithms for 2-hop cellular networks