Refine
Year of publication
Document Type
- Preprint (759)
- Article (402)
- Working Paper (119)
- Doctoral Thesis (93)
- Diploma Thesis (47)
- Conference Proceeding (41)
- Book (37)
- Bachelor Thesis (36)
- diplomthesis (28)
- Report (25)
Has Fulltext
- yes (1619)
Is part of the Bibliography
- no (1619)
Keywords
Institute
- Informatik (1619) (remove)
The specific temporal evolution of bacterial and phage population sizes, in particular bacterial depletion and the emergence of a resistant bacterial population, can be seen as a kinetic fingerprint that depends on the manifold interactions of the specific phage–host pair during the course of infection. We have elaborated such a kinetic fingerprint for a human urinary tract Klebsiella pneumoniae isolate and its phage vB_KpnP_Lessing by a modeling approach based on data from in vitro co-culture. We found a faster depletion of the initially sensitive bacterial population than expected from simple mass action kinetics. A possible explanation for the rapid decline of the bacterial population is a synergistic interaction of phages which can be a favorable feature for phage therapies. In addition to this interaction characteristic, analysis of the kinetic fingerprint of this bacteria and phage combination revealed several relevant aspects of their population dynamics: A reduction of the bacterial concentration can be achieved only at high multiplicity of infection whereas bacterial extinction is hardly accomplished. Furthermore the binding affinity of the phage to bacteria is identified as one of the most crucial parameters for the reduction of the bacterial population size. Thus, kinetic fingerprinting can be used to infer phage–host interactions and to explore emergent dynamics which facilitates a rational design of phage therapies.
Software updates are a critical success factor in mobile app ecosystems. Through publishing regular updates, platform providers enhance their operating systems for the benefit of both end users and third-party developers. It is also a way of attracting new customers. However, this platform evolution poses the risk of inadvertently introducing software problems, which can severely disturb the ecosystem’s balance by compromising its foundational technologies. So far, little to no research has addressed this issue from a user-centered perspective. The thesis at hand draws on IS post-adoption literature to investigate the potential negative influences of operating system updates on mobile app users. The release of Apple’s iOS 13 update serves as research object. Based on over half a million user reviews from the AppStore, data mining techniques are applied to study the impact of the new platform version. The results show that iOS 13 caused complications with a large number of popular apps, leading to a significant decline in user ratings and an uptrend in negative sentiment. Feature requests, functional complaints, and device compatibility are identified as the three major issue categories. These issue types are compared in terms of their quantifiable negative effect on users’ continuance intention. In essence, the findings contribute to IS research on post-adoption behavior and provide guidance to ecosystem participants in dealing with update-induced platform issues.
BIOfid is a specialized information service currently being developed to mobilize biodiversity data dormant in printed historical and modern literature and to offer a platform for open access journals on the science of biodiversity. Our team of librarians, computer scientists and biologists produce high-quality text digitizations, develop new text-mining tools and generate detailed ontologies enabling semantic text analysis and semantic search by means of user-specific queries. In a pilot project we focus on German publications on the distribution and ecology of vascular plants, birds, moths and butterflies extending back to the Linnaeus period about 250 years ago. The three organism groups have been selected according to current demands of the relevant research community in Germany. The text corpus defined for this purpose comprises over 400 volumes with more than 100,000 pages to be digitized and will be complemented by journals from other digitization projects, copyright-free and project-related literature. With TextImager (Natural Language Processing & Text Visualization) and TextAnnotator (Discourse Semantic Annotation) we have already extended and launched tools that focus on the text-analytical section of our project. Furthermore, taxonomic and anatomical ontologies elaborated by us for the taxa prioritized by the project’s target group - German institutions and scientists active in biodiversity research - are constantly improved and expanded to maximize scientific data output. Our poster describes the general workflow of our project ranging from literature acquisition via software development, to data availability on the BIOfid web portal (http://biofid.de/), and the implementation into existing platforms which serve to promote global accessibility of biodiversity data.
Risk evaluations for agricultural chemicals are necessary to preserve healthy populations of honey bee colonies. Field studies on whole colonies are limited in behavioural research, while results from lab studies allow only restricted conclusions on whole colony impacts. Methods for automated long-term investigations of behaviours within comb cells, such as brood care, were hitherto missing. In the present study, we demonstrate an innovative video method that enables within-cell analysis in honey bee (Apis mellifera) observation hives to detect chronic sublethal neonicotinoid effects of clothianidin (1 and 10 ppb) and thiacloprid (200 ppb) on worker behaviour and development. In May and June, colonies which were fed 10 ppb clothianidin and 200 ppb thiacloprid in syrup over three weeks showed reduced feeding visits and duration throughout various larval development days (LDDs). On LDD 6 (capping day) total feeding duration did not differ between treatments. Behavioural adaptation was exhibited by nurses in the treatment groups in response to retarded larval development by increasing the overall feeding timespan. Using our machine learning algorithm, we demonstrate a novel method for detecting behaviours in an intact hive that can be applied in a versatile manner to conduct impact analyses of chemicals, pests and other stressors.
Measurement of ϒ(1S) elliptic flow at forward rapidity in Pb-Pb collisions at √sNN = 5.02 TeV
(2019)
The first measurement of the ϒ(1S) elliptic flow coefficient (v2) is performed at forward rapidity (2.5 < y < 4) in Pb–Pb collisions at √sNN = 5.02 TeV with the ALICE detector at the LHC. The results are obtained with the scalar product method and are reported as a function of transverse momentum (pT) up to 15 GeV/c in the 5%–60% centrality interval. The measured Υ(1S)v2 is consistent with 0 and with the small positive values predicted by transport models within uncertainties. The v2 coefficient in 2 < pT < 15 GeV/c is lower than that of inclusive J/ψ mesons in the same pT interval by 2.6 standard deviations. These results, combined with earlier suppression measurements, are in agreement with a scenario in which the Υ(1S) production in Pb–Pb collisions at LHC energies is dominated by dissociation limited to the early stage of the collision, whereas in the J/ψ case there is substantial experimental evidence of an additional regeneration component.
Dancing is an activity that positively enhances the mood of people that consists of feeling the music and expressing it in rhythmic movements with the body. Learning how to dance can be challenging because it requires proper coordination and understanding of rhythm and beat. In this paper, we present the first implementation of the Dancing Coach (DC), a generic system designed to support the practice of dancing steps, which in its current state supports the practice of basic salsa dancing steps. However, the DC has been designed to allow the addition of more dance styles. We also present the first user evaluation of the DC, which consists of user tests with 25 participants. Results from the user test show that participants stated they had learned the basic salsa dancing steps, to move to the beat and body coordination in a fun way. Results also point out some direction on how to improve the future versions of the DC.
Iconographic representations on ancient artifacts are described in many existing databases and literature as human readable text. We applied Natural Language Processing (NLP) approaches in order to extract the semantics out of these textual descriptions and in this way enable semantic searches over them. This allows more sophisticated requests compared to the common existing keyword searches. As we show in our experiments based on numismatic datasets, the approach is generic in the sense that once the system is trained on one dataset, it can be applied without any further manual work also to datasets that have similar content. Of course, additional adaptions would further improve the results. Since the approach requires manual work only during the training phase, it can easily be applied to huge datasets without manual work and therefore without major extra costs. In fact, in our experience bigger datasets generate even better results because there is more data for training. Since our approach is not bound to a certain domain and the numismatic datasets are just an example, it could serve as a blueprint for many other areas. It could also help to build bridges between disciplines since textual iconographic descriptions are to be found also for pottery, sculpture and elsewhere.
This paper is a contribution to exploring and analyzing space-improvements in concurrent programming languages, in particular in the functional process-calculus CHF. Space-improvements are defined as a generalization of the corresponding notion in deterministic pure functional languages. The main part of the paper is the O(n ·logn) algorithm SPOPTN for offline space optimization of several parallel independent processes. Applications of this algorithm are: (i) affirmation of space improving transformations for particular classes of program transformations; (ii) support of an interpreter-based method for refuting space-improvements; and (iii) as a stand-alone offline-optimizer for space (or similar resources) of parallel processes.
Correction to: Scientifc Reports https://doi.org/10.1038/s41598-019-43857-5, published online 17 May 2019. In the original version of this Article, Jan-Hendrik Trösemeier was incorrectly affiliated with ‘Division of Allergology, Paul Ehrlich Institut, Langen, Germany’. Te correct afliations are listed below...
Gegenstand der hier vorgestellten Arbeit ist eine Applikation für die virtuelle Realität (VR), die in der Lage ist, die Struktur eines beliebigen Textes als begehbare, interaktive Stadt zu visualisieren. Darüber hinaus bietet das Programm eine besondere Textsuche an, die so in anderen konventionellen Textverarbeitungsprogrammen nicht vorzufinden ist. Dank der strukturellen Analyse und der Verwendung einiger außergewöhnlicher Analysetools des TextImager [2], ermöglicht text2City nicht nur die Suche nach bestimmten Textmustern, sondern zum Beispiel auch die Bestimmung der Textebene (Wort, Satz, Absatz, etc.) und einiges mehr. Ein weiteres Feature ist die Kommunikationsverbindung zwischen dem TextAnnotator-Service [1] und text2City, die dem Benutzer die Möglichkeit zum Annotieren bietet, aber auch von anderen Personen durchgeführte Annotationen sofort sichtbar machen kann. Für die Ausführung des Programms ist eine der beiden VRBrillen, Oculus Rift oder HTC Vive, ein für VR geeigneter PC, sowie die Software Unity nötig.
Though the range of invariance in recognition of novel objects is a basic aspect of human vision, its characterization has remained surprisingly elusive. Here we report tolerance to scale and position changes in one-shot learning by measuring recognition accuracy of Korean letters presented in a flash to non-Korean subjects who had no previous experience with Korean letters. We found that humans have significant scale-invariance after only a single exposure to a novel object. The range of translation-invariance is limited, depending on the size and position of presented objects. To understand the underlying brain computation associated with the invariance properties, we compared experimental data with computational modeling results. Our results suggest that to explain invariant recognition of objects by humans, neural network models should explicitly incorporate built-in scale-invariance, by encoding different scale channels as well as eccentricity-dependent representations captured by neurons’ receptive field sizes and sampling density that change with eccentricity. Our psychophysical experiments and related simulations strongly suggest that the human visual system uses a computational strategy that differs in some key aspects from current deep learning architectures, being more data efficient and relying more critically on eye-movements.
In pathology, tissue images are evaluated using a light microscope, relying on the expertise and experience of pathologists. There is a great need for computational methods to quantify and standardize histological observations. Computational quantification methods become more and more essential to evaluate tissue images. In particular, the distribution of tumor cells and their microenvironment are of special interest. Here, we systematically investigated tumor cell properties and their spatial neighborhood relations by a new application of statistical analysis to whole slide images of Hodgkin lymphoma, a tumor arising in lymph nodes, and inflammation of lymph nodes called lymphadenitis. We considered properties of more than 400, 000 immunohistochemically stained, CD30-positive cells in 35 whole slide images of tissue sections from subtypes of the classical Hodgkin lymphoma, nodular sclerosis and mixed cellularity, as well as from lymphadenitis. We found that cells of specific morphology exhibited significant favored and unfavored spatial neighborhood relations of cells in dependence of their morphology. This information is important to evaluate differences between Hodgkin lymph nodes infiltrated by tumor cells (Hodgkin lymphoma) and inflamed lymph nodes, concerning the neighborhood relations of cells and the sizes of cells. The quantification of neighborhood relations revealed new insights of relations of CD30-positive cells in different diagnosis cases. The approach is general and can easily be applied to whole slide image analysis of other tumor types.
The morphology of presynaptic specializations can vary greatly ranging from classical single-release-site boutons in the central nervous system to boutons of various sizes harboring multiple vesicle release sites. Multi-release-site boutons can be found in several neural contexts, for example at the neuromuscular junction (NMJ) of body wall muscles of Drosophila larvae. These NMJs are built by two motor neurons forming two types of glutamatergic multi-release-site boutons with two typical diameters. However, it is unknown why these distinct nerve terminal configurations are used on the same postsynaptic muscle fiber. To systematically dissect the biophysical properties of these boutons we developed a full three-dimensional model of such boutons, their release sites and transmitter-harboring vesicles and analyzed the local vesicle dynamics of various configurations during stimulation. Here we show that the rate of transmission of a bouton is primarily limited by diffusion-based vesicle movements and that the probability of vesicle release and the size of a bouton affect bouton-performance in distinct temporal domains allowing for an optimal transmission of the neural signals at different time scales. A comparison of our in silico simulations with in vivo recordings of the natural motor pattern of both neurons revealed that the bouton properties resemble a well-tuned cooperation of the parameters release probability and bouton size, enabling a reliable transmission of the prevailing firing-pattern at diffusion-limited boutons. Our findings indicate that the prevailing firing-pattern of a neuron may determine the physiological and morphological parameters required for its synaptic terminals.
Relying on the theory of Saward (2010) and Disch (2015), we study political representation through the lens of representative claim-making. We identify a gap between the theoretical concept of claim-making and the empirical (quantitative) assessment of representative claims made in the real world’s representative contexts. Therefore, we develop a new approach to map and quantify representative claims in order to subsequently measure the reception and validation of the claims by the audience. To test our method, we analyse all the debates of the German parliament concerned with the introduction of the gender quota in German supervisory boards from 2013 to 2017 in a two-step process. At first, we assess which constituencies the MPs claim to represent and how they justify their stance. Drawing on multiple correspondence analysis, we identify different claim patterns. Second, making use of natural language processing techniques and logistic regression on social media data, we measure if and how the asserted claims in the parliamentary debates are received and validated by the respective audience. We come to the conclusion that the constituency as ultimate judge of legitimacy has not been comprehensively conceptualized yet.
The formulation of the Partial Information Decomposition (PID) framework by Williams and Beer in 2010 attracted a significant amount of attention to the problem of defining redundant (or shared), unique and synergistic (or complementary) components of mutual information that a set of source variables provides about a target. This attention resulted in a number of measures proposed to capture these concepts, theoretical investigations into such measures, and applications to empirical data (in particular to datasets from neuroscience). In this Special Issue on “Information Decomposition of Target Effects from Multi-Source Interactions” at Entropy, we have gathered current work on such information decomposition approaches from many of the leading research groups in the field. We begin our editorial by providing the reader with a review of previous information decomposition research, including an overview of the variety of measures proposed, how they have been interpreted and applied to empirical investigations. We then introduce the articles included in the special issue one by one, providing a similar categorisation of these articles into: i. proposals of new measures; ii. theoretical investigations into properties and interpretations of such approaches, and iii. applications of these measures in empirical studies. We finish by providing an outlook on the future of the field.
This paper describes work on the morphological and syntactic annotation of Sumerian cuneiform as a model for low resource languages in general. Cuneiform texts are invaluable sources for the study of history, languages, economy, and cultures of Ancient Mesopotamia and its surrounding regions. Assyriology, the discipline dedicated to their study, has vast research potential, but lacks the modern means for computational processing and analysis. Our project, Machine Translation and Automated Analysis of Cuneiform Languages, aims to fill this gap by bringing together corpus data, lexical data, linguistic annotations and object metadata. The project’s main goal is to build a pipeline for machine translation and annotation of Sumerian Ur III administrative texts. The rich and structured data is then to be made accessible in the form of (Linguistic) Linked Open Data (LLOD), which should open them to a larger research community. Our contribution is two-fold: in terms of language technology, our work represents the first attempt to develop an integrative infrastructure for the annotation of morphology and syntax on the basis of RDF technologies and LLOD resources. With respect to Assyriology, we work towards producing the first syntactically annotated corpus of Sumerian.
Anaplastic large cell lymphoma (ALCL) and classical Hodgkin lymphoma (cHL) are lymphomas that contain CD30-expressing tumor cells and have numerous pathological similarities. Whereas ALCL is usually diagnosed at an advanced stage, cHL more frequently presents with localized disease. The aim of the present study was to elucidate the mechanisms underlying the different clinical presentation of ALCL and cHL. Chemokine and chemokine receptor expression were similar in primary ALCL and cHL cases apart from the known overexpression of the chemokines CCL17 and CCL22 in the Hodgkin and Reed-Sternberg (HRS) cells of cHL. Consistent with the overexpression of these chemokines, primary cHL cases encountered a significantly denser T cell microenvironment than ALCL. Additionally to differences in the interaction with their microenvironment, cHL cell lines presented a lower and less efficient intrinsic cell motility than ALCL cell lines, as assessed by time-lapse microscopy in a collagen gel and transwell migration assays. We thus propose that the combination of impaired basal cell motility and differences in the interaction with the microenvironment hamper the dissemination of HRS cells in cHL when compared with the tumor cells of ALCL.
Our recently developed LRSX Tool implements a technique to automatically prove the correctness of program transformations in higher-order program calculi which may permit recursive let-bindings as they occur in functional programming languages. A program transformation is correct if it preserves the observational semantics of programs- In our tool the so-called diagram method is automated by combining unification, matching, and reasoning on alpha-renamings on the higher-order metalanguage, and automating induction proofs via an encoding into termination problems of term rewrite systems. We explain the techniques, we illustrate the usage of the tool, and we report on experiments.
The focus of this paper are space-improvements of programs, which are transformations that do not worsen the space requirement during evaluations. A realistic theoretical treatment must take garbage collection method into account. We investigate space improvements under the assumption of an optimal garbage collector. Such a garbage collector is not implementable, but there is an advantage: The investigations are independent of potential changes in an implementable garbage collector and our results show that the evaluation and other similar transformations are space-improvements.
Human readers have the ability to infer knowledge from text, even if that particular information is not explicitly stated. In this thesis, we address the phenomena of text-level implicit information and outline novel automated methods for its recovery.
The main focus of this work is on two types of unexpressed content that arises between sentences (implicit discourse relations) and within sentences (implicit semantic roles).
Traditional approaches mostly rely on costly rich linguistic features, e.g., sentiment or frame-based lexicons, and require heuristics or manual feature engineering.
As an improvement, we propose a collection of generic resource-lean methods, implemented in the form of statistical background knowledge or by means of neural architectures.
Our models are largely language-independent and produce state-of-the-art performance, e.g., in the classification of Chinese implicit discourse relations, or the detection of locally covert predicative arguments in free texts.
In novel experiments, we quantitatively demonstrate that both types of implicit information are mutually dependent insofar as, for instance, some implicit roles directly correlate with implicit discourse relations of similar properties.
We show that implicit information processing further benefits downstream applications and demonstrate its applicability to the higher-level task of narrative story understanding.
In the conclusion of the dissertation, we argue for the need of implicit information processing in order to realize the goal of true natural language understanding.
The development of multimodal sensor-based applications designed to support learners with the improvement of their skills is expensive since most of these applications are tailor-made and built from scratch. In this paper, we show how the Presentation Trainer (PT), a multimodal sensor-based application designed to support the development of public speaking skills, can be modularly extended with a Virtual Reality real-time feedback module (VR module), which makes usage of the PT more immersive and comprehensive. The described study consists of a formative evaluation and has two main objectives. Firstly, a technical objective is concerned with the feasibility of extending the PT with an immersive VR Module. Secondly, a user experience objective focuses on the level of satisfaction of interacting with the VR extended PT. To study these objectives, we conducted user tests with 20 participants. Results from our test show the feasibility of modularly extending existing multimodal sensor-based applications, and in terms of learning and user experience, results indicate a positive attitude of the participants towards using the application (PT+VR module).
Programmable hardware in the form of FPGAs found its place in various high energy physics experiments over the past few decades. These devices provide highly parallel and fully configurable data transport, data formatting, and data processing capabilities with custom interfaces, even in rigid or constrained environments. Additionally, FPGA functionalities and the number of their logic resources have grown exponentially in the last few years, making FPGAs more and more suitable for complex data processing tasks. ALICE is one of the four main experiments at the LHC and specialized in the study of heavy-ion collisions. The readout chain of the ALICE detectors makes use of FPGAs at various places. The Read-Out Receiver Cards (RORCs) are one example of FPGA-based readout hardware, building the interface between the custom detector electronics and the commercial server nodes in the data processing clusters of the Data Acquisition (DAQ) system as well as the High Level Trigger (HLT). These boards are implemented as server plug-in cards with serial optical links towards the detectors. Experimental data is received via more than 500 optical links, already partly pre-processed in the FPGAs, and pushed towards the host machines. Computer clusters consisting of a few hundred nodes collect, aggregate, compress, reconstruct, and prepare the experimental data for permanent storage and later analysis. With the end of the first LHC run period in 2012 and the start of Run 2 in 2015, the DAQ and HLT systems were renewed and several detector components were upgraded for higher data rates and event rates. Increased detector link rates and obsolete host interfaces rendered it impossible to reuse the previous RORCs in Run 2.
This thesis describes the development, integration, and maintenance of the next generation of RORCs for ALICE in Run 2. A custom hardware platform, initially developed as a joint effort between the ALICE DAQ and HLT groups in the course of this work, found its place in the Run 2 readout systems of the ALICE and ATLAS experiments. The hardware fulfills all experiment requirements, matches its target performance, and has been running stable in the production systems since the start of Run 2. Firmware and software developments for the hardware evaluation, the design of the board, the mass production hardware tests, as well as the operation of the final board in the HLT, were carried out as part of this work. 74 boards were integrated into the HLT hardware and software infrastructure, with various firmware and software developments, to provide the main experimental data input and output interface of the HLT for Run 2. The hardware cluster finder, an FPGA-based data pre-processing core from the previous generation of RORCs, was ported to the new hardware. It has been improved and extended to meet the experimental requirements throughout Run 2. The throughput of this firmware component could be doubled and the algorithm extended, providing an improved noise rejection and an increased overall mean data compression ratio compared to its previous implementation. The hardware cluster finder forms a crucial component in the HLT data reconstruction and compression scheme with a processing performance of one board equivalent to around ten server nodes for comparable processing steps in software.
The work on the firmware development, especially on the hardware cluster finder, once more demonstrated that developing and maintaining data processing algorithms with the common low-level hardware description methods is tedious and time-consuming. Therefore, a high-level synthesis (HLS) hardware description method applying dataflow computing at an algorithmic level to FPGAs was evaluated in this context. The hardware cluster finder served as an example of a typical data processing algorithm in a high energy physics readout application. The existing and highly optimized low-level implementation provided a reference for comparisons in terms of throughput and resource usage. The cluster finder algorithm could be implemented in the dataflow description with comparably little effort, providing fast development cycles, compact code and at, the same time, simplified extension and maintenance options. The performance results in terms of throughput and resource usage are comparable to the manual implementation. The dataflow environment proved to be highly valuable for design space explorations. An integration of the dataflow description into the HLT firmware and software infrastructure could be demonstrated as a proof of concept. A high-level hardware description could ease both the design space exploration, the initial development, the maintenance, and the extension of hardware algorithms for high energy physics readout applications.
Background: Microarray analysis represents a powerful way to test scientific hypotheses on the functionality of cells. The measurements consider the whole genome, and the large number of generated data requires sophisticated analysis. To date, no gold-standard for the analysis of microarray images has been established. Due to the lack of a standard approach there is a strong need to identify new processing algorithms.
Methods: We propose a novel approach based on hyperbolic partial differential equations (PDEs) for unsupervised spot segmentation. Prior to segmentation, morphological operations were applied for the identification of co-localized groups of spots. A grid alignment was performed to determine the borderlines between rows and columns of spots. PDEs were applied to detect the inflection points within each column and row; vertical and horizontal luminance profiles were evolved respectively. The inflection points of the profiles determined borderlines that confined a spot within adapted rectangular areas. A subsequent k-means clustering determined the pixels of each individual spot and its local background.
Results: We evaluated the approach for a data set of microarray images taken from the Stanford Microarray Database (SMD). The data set is based on two studies on global gene expression profiles of Arabidopsis Thaliana. We computed values for spot intensity, regression ratio, and coefficient of determination. For spots with irregular contours and inner holes, we found intensity values that were significantly different from those determined by the GenePix Pro microarray analysis software. We determined the set of differentially expressed genes from our intensities and identified more activated genes than were predicted by the GenePix software.
Conclusions: Our method represents a worthwhile alternative and complement to standard approaches used in industry and academy. We highlight the importance of our spot segmentation approach, which identified supplementary important genes, to better explains the molecular mechanisms that are activated in a defense responses to virus and pathogen infection.
The production of K∗(892)0 and ϕ(1020) mesons has been measured in p–Pb collisions at √sNN = 5.02 TeV. K∗0 and ϕ are reconstructed via their decay into charged hadrons with the ALICE detector in the rapidity range - 0.5 < y < 0. The transverse momentum spectra, measured as a function of the multiplicity, have a pT range from 0 to 15 GeV/c for K∗0 and from 0.3 to 21 GeV/c for ϕ. Integrated yields, mean transverse momenta and particle ratios are reported and compared with results in pp collisions at √s = 7 TeV and Pb–Pb collisions at √sNN = 2.76 TeV. In Pb–Pb and p–Pb collisions, K∗0 and ϕ probe the hadronic phase of the system and contribute to the study of particle formation mechanisms by comparison with other identified hadrons. For this purpose, the mean transverse momenta and the differential proton-to-ϕ ratio are discussed as a function of the multiplicity of the event. The short-lived K∗0 is measured to investigate re-scattering effects, believed to be related to the size of the system and to the lifetime of the hadronic phase.
Heterologously expressed genes require adaptation to the host organism to ensure adequate levels of protein synthesis, which is typically approached by replacing codons by the target organism’s preferred codons. In view of frequently encountered suboptimal outcomes we introduce the codon-specific elongation model (COSEM) as an alternative concept. COSEM simulates ribosome dynamics during mRNA translation and informs about protein synthesis rates per mRNA in an organism- and context-dependent way. Protein synthesis rates from COSEM are integrated with further relevant covariates such as translation accuracy into a protein expression score that we use for codon optimization. The scoring algorithm further enables fine-tuning of protein expression including deoptimization and is implemented in the software OCTOPOS. The protein expression score produces competitive predictions on proteomic data from prokaryotic, eukaryotic, and human expression systems. In addition, we optimized and tested heterologous expression of manA and ova genes in Salmonella enterica serovar Typhimurium. Superiority over standard methodology was demonstrated by a threefold increase in protein yield compared to wildtype and commercially optimized sequences.
Eingebettete Systeme sind Rechnersysteme, die in einem technischen Umfeld eingebettet sind und dort ihre Arbeit verrichten. Kennzeichen heutiger und zukünftiger eingebetteter Systeme sind, dass sie in einer immer größeren Anzahl in der Industrie, im Haushalt und in Büros, in Eisenbahnen und Flugzeugen und in vielen weiteren Umgebungen auftreten. Sie sind oftmals stark vernetzt und müssen hochverlässlich sein, um Unfälle zu vermeiden und so Anwender und Nutzer vor Schaden zu bewahren. Die Beherrschung dieser eingebetteten Systeme ist meist hochkomplex, da durch die Vernetzung eine Vielzahl von Komponenten zusammenarbeiten. Für den Anwender ist es daher schwer, den Überblick zu behalten. Im Hinblick auf die Verlässlichkeit ist es wichtig, Reaktionen auf Fehler und unvorhergesehene Situationen in diesen Systemen innerhalb definierter Zeitschranken zu liefern, um Schaden zu vermeiden.
Selbstorganisation wird heutzutage als probates Mittel angesehen, um die Herausforderungen, die sich mit der Inbetriebnahme, Nutzung und Instandhaltung von komplexen eingebetteten Systemen ergeben, zu meistern. Der Beitrag dieser Arbeit ist eine Untersuchung selbstorganisierender eingebetteter Systeme:
Im ersten Teil wird ein Überblick über den aktuellen Stand der Forschung bei eingebetteten Systemen sowie über den Bereich der Selbstorganisation für eingebettete Systeme gegeben. Dabei wird die Idee des Organic Computings beschrieben, welches sich mit Selbstorganisationsprinzipien in IT-Systemen beschäftigt, und es werden aktuelle Forschungstrends dazu beschrieben.
Im zweiten Teil der Arbeit werden eigene Arbeiten im Feld von selbstorganisierenden eingebetteten Systemen vorgestellt. Sie behandeln verschiedene Aspekte eines künstlichen Hormonsystems (KHS), welches zur selbstorganisierten Verteilung von Tasks auf einer Menge von vernetzten Prozessoren genutzt werden kann. Dabei werden einerseits grundlegende Definitionen des Organic Computings im Bezug auf das KHS untersucht und bewertet. Andererseits werden neue Lerntechniken für das KHS untersucht, die sich am maschinellen Lernen orientieren. Außerdem wird ein mehrstufiges KHS entwickelt und evaluiert, um die Vergabe einer sehr großen Anzahl von Tasks (≥ 1000) auf einer sehr großen Anzahl von Prozessoren (≥ 10000) zu ermöglichen.
CRFVoter : gene and protein related object recognition using a conglomerate of CRF-based tools
(2019)
Background: Gene and protein related objects are an important class of entities in biomedical research, whose identification and extraction from scientific articles is attracting increasing interest. In this work, we describe an approach to the BioCreative V.5 challenge regarding the recognition and classification of gene and protein related objects. For this purpose, we transform the task as posed by BioCreative V.5 into a sequence labeling problem. We present a series of sequence labeling systems that we used and adapted in our experiments for solving this task. Our experiments show how to optimize the hyperparameters of the classifiers involved. To this end, we utilize various algorithms for hyperparameter optimization. Finally, we present CRFVoter, a two-stage application of Conditional Random Field (CRF) that integrates the optimized sequence labelers from our study into one ensemble classifier.
Results: We analyze the impact of hyperparameter optimization regarding named entity recognition in biomedical research and show that this optimization results in a performance increase of up to 60%. In our evaluation, our ensemble classifier based on multiple sequence labelers, called CRFVoter, outperforms each individual extractor’s performance. For the blinded test set provided by the BioCreative organizers, CRFVoter achieves an F-score of 75%, a recall of 71% and a precision of 80%. For the GPRO type 1 evaluation, CRFVoter achieves an F-Score of 73%, a recall of 70% and achieved the best precision (77%) among all task participants.
Conclusion: CRFVoter is effective when multiple sequence labeling systems are to be used and performs better then the individual systems collected by it.
We explore space improvements in LRP, a polymorphically typed call-by-need functional core language. A relaxed space measure is chosen for the maximal size usage during an evaluation. It Abstracts from the details of the implementation via abstract machines, but it takes garbage collection into account and thus can be seen as a realistic approximation of space usage. The results are: a context lemma for space improving translations and for space equivalences; all but one reduction rule of the calculus are shown to be space improvements, and the exceptional one, the copy-rule, is shown to increase space only moderately.
Several further program transformations are shown to be space improvements or space equivalences, in particular the translation into machine expressions is a space equivalence. These results are a step Forward in making predictions about the change in runtime space behavior of optimizing transformations in callbyneed functional languages.
We explore space improvements in LRP, a polymorphically typed call-by-need functional core language. A relaxed space measure is chosen for the maximal size usage during an evaluation. It Abstracts from the details of the implementation via abstract machines, but it takes garbage collection into account and thus can be seen as a realistic approximation of space usage. The results are: a context lemma for space improving translations and for space equivalences; all but one reduction rule of the calculus are shown to be space improvements, and the exceptional one, the copy-rule, is shown to increase space only moderately.
Several further program transformations are shown to be space improvements or space equivalences, in particular the translation into machine expressions is a space equivalence. These results are a step Forward in making predictions about the change in runtime space behavior of optimizing transformations in callbyneed functional languages.
The synchronous pi-calculus is translated into a core language of Concurrent Haskell extended by futures (CHF). The translation simulates the synchronous message-passing of the pi-calculus by sending messages and adding synchronization using Concurrent Haskell's mutable shared-memory locations (MVars). The semantic criterion is a contextual semantics of the pi-calculus and of CHF using may- and should-convergence as observations. The results are equivalence with respect to the observations, full abstraction of the translation of closed processes, and adequacy of the translation on open processes. The translation transports the semantics of the pi-calculus processes under rather strong criteria, since error-free programs are translated into error-free ones, and programs without non-deterministic error possibilities are also translated into programs without non-deterministic error-possibilities. This investigation shows that CHF embraces the expressive power and the concurrency capabilities of the pi-calculus.
The hepatitis C virus (HCV) RNA replication cycle is a dynamic intracellular process occurring in three-dimensional space (3D), which is difficult both to capture experimentally and to visualize conceptually. HCV-generated replication factories are housed within virus-induced intracellular structures termed membranous webs (MW), which are derived from the Endoplasmatic Reticulum (ER). Recently, we published 3D spatiotemporal resolved diffusion–reaction models of the HCV RNA replication cycle by means of surface partial differential equation (sPDE) descriptions. We distinguished between the basic components of the HCV RNA replication cycle, namely HCV RNA, non-structural viral proteins (NSPs), and a host factor. In particular, we evaluated the sPDE models upon realistic reconstructed intracellular compartments (ER/MW). In this paper, we propose a significant extension of the model based upon two additional parameters: different aggregate states of HCV RNA and NSPs, and population dynamics inspired diffusion and reaction coefficients instead of multilinear ones. The combination of both aspects enables realistic modeling of viral replication at all scales. Specifically, we describe a replication complex state consisting of HCV RNA together with a defined amount of NSPs. As a result of the combination of spatial resolution and different aggregate states, the new model mimics a cis requirement for HCV RNA replication. We used heuristic parameters for our simulations, which were run only on a subsection of the ER. Nevertheless, this was sufficient to allow the fitting of core aspects of virus reproduction, at least qualitatively. Our findings should help stimulate new model approaches and experimental directions for virology.
"Prognosen sind schwierig, besonders, wenn sie die Zukunft betreffen", sagt ein geflügeltes Wort. Die letzte Finanzkrise ist dafür ein gutes Beispiel, denn die wenigsten Analysten und Wirtschaftsweisen haben sie kommen sehen. Da Finanzkrisen glücklicherweise selten sind, ist es allerdings schwierig, Modelle zu entwickeln, die rechtzeitig vor einem Crash warnen.
Die wahrscheinlich beste Entscheidung : wie Online-Algorithmen mit der unsicheren Zukunft rechnen
(2018)
Lohnt es sich, als Skianfänger in einem schneeunsicheren Jahr Skier zu kaufen? Oder ist es günstiger, sie zu mieten? Oft müssen wir Entscheidungen treffen, ohne genügend Informationen über die Zukunft zu haben. Das gilt in noch größerem Maße für Rechnersysteme, die große Datenmengen verarbeiten und schnelle Entscheidungen treffen müssen. Damit sie trotz einer Vielzahl von Unsicherheiten erfolgreich arbeiten können, entwickeln Informatiker OnlineAlgorithmen.
Künstliche Intelligenz (KI), also intelligente Software, führt heutzutage Aufgaben aus, die man einst nur Menschen zutraute. Schon heute ist sie in vielen Bereichen unserer Gesellschaft angekommen – man denke an selbstfahrende Fahrzeuge, medizinische Diagnostik, Übersetzungsprogramme, persönliche Gesprächsassistenten, Suchfunktionen und Robotik. Doch wie weit können wir KI-Systemen vertrauen?
LSTMVoter : chemical named entity recognition using a conglomerate of sequence labeling tools
(2019)
Background: Chemical and biomedical named entity recognition (NER) is an essential preprocessing task in natural language processing. The identification and extraction of named entities from scientific articles is also attracting increasing interest in many scientific disciplines. Locating chemical named entities in the literature is an essential step in chemical text mining pipelines for identifying chemical mentions, their properties, and relations as discussed in the literature. In this work, we describe an approach to the BioCreative V.5 challenge regarding the recognition and classification of chemical named entities. For this purpose, we transform the task of NER into a sequence labeling problem. We present a series of sequence labeling systems that we used, adapted and optimized in our experiments for solving this task. To this end, we experiment with hyperparameter optimization. Finally, we present LSTMVoter, a two-stage application of recurrent neural networks that integrates the optimized sequence labelers from our study into a single ensemble classifier.
Results: We introduce LSTMVoter, a bidirectional long short-term memory (LSTM) tagger that utilizes a conditional random field layer in conjunction with attention-based feature modeling. Our approach explores information about features that is modeled by means of an attention mechanism. LSTMVoter outperforms each extractor integrated by it in a series of experiments. On the BioCreative IV chemical compound and drug name recognition (CHEMDNER) corpus, LSTMVoter achieves an F1-score of 90.04%; on the BioCreative V.5 chemical entity mention in patents corpus, it achieves an F1-score of 89.01%.
Availability and implementation: Data and code are available at https://github.com/texttechnologylab/LSTMVoter.
Mit der Smart Learning Infrastruktur wurde ein neuartiges didaktisches Konzept für Kurse in der Weiterbildung entwickelt. Diese Infrastruktur ist vielfältig anwendbar. Erste Analysen von Kursen zeigen, dass TeilnehmerInnen, die alle Übungen korrekt abgearbeitet haben, eine bessere Note erreichen als die Durchschnittsnote. Dieser Beitrag beschreibt ein Konzept für ein Gamification-Modul, welches mit spielerischen Elementen möglichst frühzeitig dazu animiert, alle Übungen eines Kurses korrekt und mit Verstand abzuarbeiten.
We propose and create a new data model for learning specific environments and learning analytics applications. This is motivated from the experience in the Fiber Bundle Data Model used for large - time and space dependent - data. Our proposed data model integrates file or stream-based data structures from capturing devices more easily. Learning analytics algorithms are added directly to the data, and formulation of queries and analytics is done in Python. It is designed to improve collaboration in the field of learning analytics. We leverage a hierarchical data structure, where varying data is located near the leaves. Abstract data types are identified in four distinct pathways, which allow storing most diverse data sources. We compare different implementations regarding its memory footprint and performance. Our tests indicate that LeAn Bundles can be smaller than a naïve xAPI export. The benchmarks show that the performance is comparable to a MongoDB, while having the benefit of being portable and extensible.
Digitale Kompetenzen von Hochschullehrenden messen : Validierungsstudie eines Kompetenzrasters
(2018)
Der Beitrag beschreibt die Entwicklung eines Kompetenzrasters zur Erfassung digitaler Kompetenzen von Hochschullehrenden und stellt Ergebnisse der Validierung des Rasters vor. Dazu werden die Ergebnisse eines Pre-Tests (N=90) unter Teilnehmenden eines E-LearningQualifizierungsangebots inferenzstatistisch ausgewertet. Zusätzlich werden zur äußeren Validierung des Kompetenzrasters Ergebnisse mit Aussagen der Befragungsteilnehmer*innen verglichen, die mit Hilfe qualitativer Methoden aus E-Portfolios gewonnen wurden. Die skalenanalytischen Befunde erbrachten für sechs der acht Subdimensionen digitaler Kompetenz eindeutige, einfaktorielle Lösungen mit guten Varianzaufklärungen. Die Subskalen verfügen über hohe interne Konsistenzen. Zwei Dimensionen trennen sich faktorenanalytisch in weitere Subtests auf, die sich im Test ebenfalls als reliabel erweisen. Zur Validität des Kompetenzrasters konnten durch Zusammenhänge mit Aussagen aus E-Portfolios positive Belege gesammelt werden.
Wie eine "Heilslehre" überzieht der Begriff "Digitalisierung" fast alle Lebensbereiche – natürlich auch den Bildungsbereich. Gerade wir Informatiker*innen sind gefordert, diese Wege der Bildungstransformation mitzugestalten. Wir zusammen mit den Erziehungswissenschaftler*innen und Psychologen*innen müssen identifizieren, aufzeigen und vorbildlich umsetzen, was sinnvoll und möglich ist. Wir sind diejenigen, die die Bedingungen des Gelingens und auch die der Irrwege erforschen und aufzeigen müssen. Digitalisierungswahnsinn brauchen wir nicht!
Die 16. Jahrestagung DeLFI 2018 der Fachgruppe eLearning der Gesellschaft für Informatik e. V. findet vom 10. bis 13.September 2018 an der Johann Wolfgang Goethe – Universität, Frankfurt am Main statt, gemeinsam mit der 8. Tagung für Hochschuldidaktik der Informatik HDI 2018. ...
The endoplasmic reticulum–mitochondria encounter structure (ERMES) connects the mitochondrial outer membrane with the ER. Multiple functions have been linked to ERMES, including maintenance of mitochondrial morphology, protein assembly and phospholipid homeostasis. Since the mitochondrial distribution and morphology protein Mdm10 is present in both ERMES and the mitochondrial sorting and assembly machinery (SAM), it is unknown how the ERMES functions are connected on a molecular level. Here we report that conserved surface areas on opposite sides of the Mdm10 β-barrel interact with SAM and ERMES, respectively. We generated point mutants to separate protein assembly (SAM) from morphology and phospholipid homeostasis (ERMES). Our study reveals that the β-barrel channel of Mdm10 serves different functions. Mdm10 promotes the biogenesis of α-helical and β-barrel proteins at SAM and functions as integral membrane anchor of ERMES, demonstrating that SAM-mediated protein assembly is distinct from ER-mitochondria contact sites.
In this paper, we study the limit of compactness which is a graph index originally introduced for measuring structural characteristics of hypermedia. Applying compactness to large scale small-world graphs (Mehler, 2008) observed its limit behaviour to be equal 1. The striking question concerning this finding was whether this limit behaviour resulted from the specifics of small-world graphs or was simply an artefact. In this paper, we determine the necessary and sufficient conditions for any sequence of connected graphs resulting in a limit value of CB = 1 which can be generalized with some consideration for the case of disconnected graph classes (Theorem 3). This result can be applied to many well-known classes of connected graphs. Here, we illustrate it by considering four examples. In fact, our proof-theoretical approach allows for quickly obtaining the limit value of compactness for many graph classes sparing computational costs.
Virtual machines are for the most part not used inside of high-energy physics (HEP) environments. Even though they provide a high degree of isolation, the performance overhead they introduce is too great for them to be used. With the rising number of container technologies and their increasing separation capabilities, HEP-environments are evaluating if they could utilize the technology. The container images are small and self-contained which allows them to be easily distributed throughout the global environment. They also offer a near native performance while at the same time aproviding an often acceptable level of isolation. Only the needed services and libraries are packed into an image and executed directly by the host kernel. This work compared the performance impact of the three container technologies Docker, rkt and Singularity. The host kernel was additionally hardened with grsecurity and PaX to strengthen its security and make an exploitation from inside a container harder. The execution time of a physics simulation was used as a benchmark. The results show that the different container technologies have a different impact on the performance. The performance loss on a stock kernel is small; in some cases they were even faster than no container. Docker showed overall the best performance on a stock kernel. The difference on a hardened kernel was bigger than on a stock kernel, but in favor of the container technologies. rkt showed performed in almost all cases better than all the others.
Unter Web-based Trainings (WBTs) versteht man multimediale, interaktive und thematisch abgeschlossene Lerneinheiten in einem Browser. Seit der Entstehung des Internets in den 1990er Jahren sind diese ein wichtiger und etablierter Baustein bei der Konzeption und Entwicklung von eLearning-Szenarien. Diese Lerneinheiten werden üblicherweise von Lehrenden mit entsprechenden Autorensystemen erstellt. In selteneren Fällen handelt es sich bei deren Umsetzungen um individuell programmierte Einzellösungen. Betrachtet man WBTs aus der Sicht der Lernenden, dann lässt sich feststellen, dass zunehmend auch nicht explizit als Lerneinheiten erstellte Inhalte genutzt werden, die jedoch genau den Bedürfnissen des jeweiligen Lernenden entsprechen (im Rahmen des informellen und selbstgesteuerten Lernens). Zum einen liegt das an der zunehmenden Verfügbarkeit und Vielfalt von „alternativen Lerninhalten“ im Internet generell (freie Lizenzen und innovative Autorentools). Zum anderen aber auch an der Möglichkeit, diese Inhalte von überall aus und zu jeder Zeit einfach finden zu können (mobiles Internet, Suchmaschinen und Sprachassistenten) bzw. eingeordnet und empfohlen zu bekommen (Empfehlungssysteme und soziale Medien).
Aus dieser Veränderung heraus ergibt sich im Rahmen dieser Dissertation die zentrale Fragestellung, ob das Konzept eines dedizierten WBT-Autorensystems den neuen Anforderungen von frei verfügbaren, interaktiven Lerninhalten (Khan Academy, YouTube und Wikipedia) und einer Vielzahl ständig wachsender und kostenfreier Autorentools für beliebige Web-Inhalte (H5P, PowToon oder Pageflow) überhaupt noch gerecht wird und wo in diesem Fall genau die Alleinstellungsmerkmale eines WBTs liegen?
Zur Beantwortung dieser Frage beschäftigt sich die Arbeit grundlegend mit dem Begriff „Web-based Training“, den über die Zeit geänderten Rahmenbedingungen und den daraus resultierenden Implikationen für die Entwicklung von WBT-Autorensystemen. Mittels des gewählten Design-based Research (DBR)-Ansatzes konnte durch kontinuierliche Zyklen von Gestaltung, Durchführung, Analyse und Re-Design am Beispiel mehrerer eLearning-Projekte der Begriff WBT neudefininiert bzw. reinterpretiert werden, so dass sich der Fokus der Definition auf das konzentriert, was WBTs im Vergleich zu anderen Inhalten und Funktionen im Internet im Kern unterscheidet: dem Lehr-/Lernaspekt (nachfolgend Web-based Training 2.0 (WBT 2.0)).
Basierend auf dieser Neudefinition konnten vier Kernfunktionalitäten ausgearbeitet werden, die die zuvor genannten Herausforderungen adressieren und in Form eines Design Frameworks detailliert beschreiben. Untersucht und entwickelt wurden die unterschiedlichen Aspekte und Funktionen der WBTs 2.0 anhand der iterativen „Meso-Zyklen“ des DBR-Ansatzes, wobei jedes der darin durchgeführten Projekte auch eigene Ergebnisse mit sich bringt, welche jeweils unter didaktischen und vor allem aber technischen Gesichtspunkten erörtert wurden. Die dadurch gewonnenen Erkenntnisse flossen jeweils in den Entwicklungsprozess der LernBar ein („Makro-Zyklus“), ein im Rahmen dieser Arbeit und von studiumdigitale, der zentralen eLearning-Einrichtung der Goethe-Universität, entwickeltes WBT-Autorensystem. Dabei wurden die Entwicklungen kontinuierlich unter Einbezug von Nutzerfeedbacks (jährliche Anwendertreffen, Schulungen, Befragungen, Support) überprüft und weiterentwickelt.
Abschließend endet der letzte Entwicklungszyklus des DBR-Ansatzes mit der Konzeption und Umsetzung von drei WBT 2.0-Systemkomponenten, wodurch sich flexibel beliebige Web-Inhalte mit entsprechenden WBT 2.0-Funktionalitäten erweitern lassen, um auch im Kontext von offenen Lehr-/Lernprozessen durchgeführte Aktivitäten transparent, nachvollziehbar und somit überprüfbar zu machen (Constructive Alignment).
Somit bietet diese Forschungsarbeit einen interdisziplinären, nutzerzentrierten und in der Praxis erprobten Ansatz für die Umsetzung und den Einsatz von WBTs im Kontext offener Lehr-/Lernprozesse. Dabei verschiebt sich der bisherige Fokus von der reinen Medienproduktion hin zu einem ganzheitlichen Ansatz, bei dem der Lehr-/Lernaspekt im Vordergrund steht (Lernbedarf erkennen, decken und überprüfen). Entscheidend ist dabei, dass zum Decken eines Lernbedarfs sämtliche zur Verfügung stehenden Ressourcen des Internets genutzt werden können, wobei WBTs 2.0 dazu lediglich den didaktischen Prozess definieren und diesen für die Lehrenden und Lernende transparent und zugänglich machen.
WBTs 2.0 profitieren dadurch zukünftig von der zunehmenden Vielfalt und Verfügbarkeit von Inhalten und Funktionen im Internet und ermöglichen es, den Entwicklern von WBT 2.0-Autorensystemen sich auf das Wesentliche zu konzentrieren: den Lehr-/Lernprozess.
Die vorliegende Arbeit lässt sich in den Bereich Data Science einordnen. Data Science verwendet Verfahren aus dem Bereich Computer Science, Algorithmen aus der Mathematik und Statistik sowie Domänenwissen, um große Datenmengen zu analysieren und neue Erkenntnisse zu gewinnen. In dieser Arbeit werden verschiedene Forschungsbereiche aus diesen verwendet. Diese umfassen die Datenanalyse im Bereich von Big Data (soziale Netzwerke, Kurznachrichten von Twitter), Opinion Mining (Analyse von Meinungen auf Basis eines Lexikons mit meinungstragenden Phrasen) sowie Topic Detection (Themenerkennung)....
Ergebnis 1: Sentiment Phrase List (SePL)
Im Forschungsbereich Opinion Mining spielen Listen meinungstragender Wörter eine wesentliche Rolle bei der Analyse von Meinungsäußerungen. Das im Rahmen dieser Arbeit entwickelte Vorgehen zur automatisierten Generierung einer solchen Liste leistet einen wichtigen Forschungsbeitrag in diesem Gebiet. Der neuartige Ansatz ermöglicht es einerseits, dass auch Phrasen aus mehreren Wörtern (inkl. Negationen, Verstärkungs- und Abschwächungspartikeln) sowie Redewendungen enthalten sind, andererseits werden die Meinungswerte aller Phrasen auf Basis eines entsprechenden Korpus automatisiert berechnet. Die Sentiment Phrase List sowie das Vorgehen wurden veröffentlicht und können von der Forschungsgemeinde genutzt werden [121, 123]. Die Erstellung basiert auf einer textuellen sowie zusätzlich numerischen Bewertung, welche typischerweise in Kundenrezensionen verwendet werden (beispielsweise der Titel und die Sternebewertung bei Amazon Kundenrezensionen). Es können weitere Datenquellen verwendet werden, die eine derartige Bewertung aufweisen. Auf Basis von ca. 1,5 Millionen deutschen Kundenrezensionen wurden verschiedene Versionen der SePL erstellt und veröffentlicht [120].
Ergebnis 2: Algorithmus auf Basis der SePL
Mit Hilfe der SePL und den darin enthaltenen meinungstragenden Phrasen ergeben sich Verbesserungen für lexikonbasierte Verfahren bei der Analyse von Meinungsäußerungen. Phrasen werden im Text häufig durch andere Wörter getrennt, wodurch eine Identifizierung der Phrasen erforderlich ist. Der Algorithmus für eine lexikonbasierte Meinungsanalyse wurde veröffentlicht [176]. Er basiert auf meinungstragenden Phrasen bestehend aus einem oder mehreren Wörtern. Da für einzelne Phrasen unterschiedliche Meinungswerte vorliegen, ist eine genauere Bewertung als mit bisherigen Ansätzen möglich. Dies ermöglicht, dass meinungstragende Phrasen aus dem Text extrahiert und anhand der in der SePL enthaltenen Einträge differenziert bewertet werden können. Bisherige Ansätze nutzen häufig einzelne meinungstragende Wörter. Der Meinungswert für beispielsweise eine Verneinung muss nicht anhand eines generellen Vorgehens erfolgen. In aktuellen Verfahren wird der Wert eines meinungstragenden Wortes bei Vorhandensein einer Verneinung bisher meist invertiert, was häufig falsche Ergebnisse liefert. Die Liste enthält im besten Fall sowohl einen Meinungswert für das einzelne Wort und seine Verneinung (z.B. „schön“ und „nicht schön“).
1.3 übersicht der hauptergebnisse 5
Ergebnis 3: Evaluierung der Anwendung der SePL
Der Algorithmus aus Ergebnis 2 wurde mit Rezensionen der Bewertungsplattform CiaoausdemBereichderAutomobilversicherunge valuiert.Dabei wurden wesentliche Fehlerquellen aufgezeigt [176], die entsprechende Verbesserungen ermöglichen. Weiterhin wurde mit der SePL eine Evaluation anhand eines Maschinenlernverfahrens auf Basis einer Support Vector Machine durchgeführt. Hierbei wurden verschiedene bestehende lexikalische Ressourcen mit der SePL verglichen sowie deren Einsatz in verschiedenen Domänen untersucht. Die Ergebnisse wurden in [115] veröffentlicht.
Ergebnis 4: Forschungsprojekt PoliTwi - Themenerkennung politischer Top-Themen
Mit dem Forschungsprojekt PoliTwi wurden einerseits die erforderlichen Daten von Twitter gesammelt. Andererseits werden der breiten Öffentlichkeit fortlaufend aktuelle politische Top-Themen über verschiedene Kanäle zur Verfügung gestellt. Für die Evaluation der angestrebten Verbesserungen im Bereich der Themenerkennung in Verbindung mit einer Meinungsanalyse liegen die erforderlichen Daten über einen Zeitraum von bisher drei Jahren aus der Domäne Politik vor. Auf Basis dieser Daten konnte die Themenerkennung durchgeführt werden. Die berechneten Themen wurden mit anderen Systemen wie Google Trends oder Tagesschau Meta verglichen (siehe Kapitel 5.3). Es konnte gezeigt werden, dass die Meinungsanalyse die Themenerkennung verbessern kann. Die Ergebnisse des Projekts wurden in [124] veröffentlicht. Der Öffentlichkeit und insbesondere Journalisten und Politikern wird zudem ein Service (u.a. anhand des Twitter-Kanals unter https://twitter.com/politwi) zur Verfügung gestellt, anhand dessen sie über aktuelle Top-Themen informiert werden. Nachrichtenportale wie FOCUS Online nutzten diesen Service bei ihrer Berichterstattung (siehe Kapitel 4.3.6.1). Die Top-Themen werden seit Mitte 2013 ermittelt und können zudem auf der Projektwebseite [119] abgerufen werden.
Ergebnis 5: Erweiterung lexikalischer Ressourcen auf Konzeptebene
Das noch junge Forschungsgebiet des Concept-level Sentiment Analysis versucht bisherige Ansätze der Meinungsanalyse dadurch zu verbessern, dass Meinungsäußerungen auf Konzeptebene analysiert werden. Eine Voraussetzung sind Listen meinungstragender Wörter, welche differenzierte Betrachtungen anhand unterschiedlicher Kontexte ermöglichen. Anhand der Top-Themen und deren Kontext wurde ein Vorgehen entwickelt, welches die Erstellung bzw. Ergänzung dieser Listen ermöglicht. Es wurde gezeigt, wie Meinungen in unterschiedlichen Kontexten differenziert bewertet werden und diese Information in lexikalischen Ressourcen aufgenommen werden können, was im Bereich der Concept-level Sentiment Analysis genutzt werden kann. Das Vorgehen wurde in [124] veröffentlicht.
Students of computer science studies enter university education with very different competencies, experience and knowledge. 145 datasets collected of freshmen computer science students by learning management systems in relation to exam outcomes and learning dispositions data (e. g. student dispositions, previous experiences and attitudes measured through self-reported surveys) has been exploited to identify indicators as predictors of academic success and hence make effective interventions to deal with an extremely heterogeneous group of students.