Refine
Year of publication
Document Type
- Monograph/Edited Volume (1210) (remove)
Language
- English (1210) (remove)
Keywords
Institute
- Institut für Mathematik (353)
- Wirtschaftswissenschaften (191)
- Hasso-Plattner-Institut für Digital Engineering gGmbH (89)
- Institut für Informatik und Computational Science (82)
- Department Linguistik (47)
- Institut für Anglistik und Amerikanistik (46)
- Sozialwissenschaften (39)
- Institut für Physik und Astronomie (35)
- Hasso-Plattner-Institut für Digital Engineering GmbH (33)
- Department Psychologie (25)
In recent years, computer vision algorithms based on machine learning have seen rapid development. In the past, research mostly focused on solving computer vision problems such as image classification or object detection on images displaying natural scenes. Nowadays other fields such as the field of cultural heritage, where an abundance of data is available, also get into the focus of research. In the line of current research endeavours, we collaborated with the Getty Research Institute which provided us with a challenging dataset, containing images of paintings and drawings. In this technical report, we present the results of the seminar "Deep Learning for Computer Vision". In this seminar, students of the Hasso Plattner Institute evaluated state-of-the-art approaches for image classification, object detection and image recognition on the dataset of the Getty Research Institute. The main challenge when applying modern computer vision methods to the available data is the availability of annotated training data, as the dataset provided by the Getty Research Institute does not contain a sufficient amount of annotated samples for the training of deep neural networks. However, throughout the report we show that it is possible to achieve satisfying to very good results, when using further publicly available datasets, such as the WikiArt dataset, for the training of machine learning models.
Data dependencies, or integrity constraints, are used to improve the quality of a database schema, to optimize queries, and to ensure consistency in a database. In the last years conditional dependencies have been introduced to analyze and improve data quality. In short, a conditional dependency is a dependency with a limited scope defined by conditions over one or more attributes. Only the matching part of the instance must adhere to the dependency. In this paper we focus on conditional inclusion dependencies (CINDs). We generalize the definition of CINDs, distinguishing covering and completeness conditions. We present a new use case for such CINDs showing their value for solving complex data quality tasks. Further, we define quality measures for conditions inspired by precision and recall. We propose efficient algorithms that identify covering and completeness conditions conforming to given quality thresholds. Our algorithms choose not only the condition values but also the condition attributes automatically. Finally, we show that our approach efficiently provides meaningful and helpful results for our use case.
Data obtained from foreign data sources often come with only superficial structural information, such as relation names and attribute names. Other types of metadata that are important for effective integration and meaningful querying of such data sets are missing. In particular, relationships among attributes, such as foreign keys, are crucial metadata for understanding the structure of an unknown database. The discovery of such relationships is difficult, because in principle for each pair of attributes in the database each pair of data values must be compared. A precondition for a foreign key is an inclusion dependency (IND) between the key and the foreign key attributes. We present with Spider an algorithm that efficiently finds all INDs in a given relational database. It leverages the sorting facilities of DBMS but performs the actual comparisons outside of the database to save computation. Spider analyzes very large databases up to an order of magnitude faster than previous approaches. We also evaluate in detail the effectiveness of several heuristics to reduce the number of necessary comparisons. Furthermore, we generalize Spider to find composite INDs covering multiple attributes, and partial INDs, which are true INDs for all but a certain number of values. This last type is particularly relevant when integrating dirty data as is often the case in the life sciences domain - our driving motivation.
Cyber-physical systems achieve sophisticated system behavior exploring the tight interconnection of physical coupling present in classical engineering systems and information technology based coupling. A particular challenging case are systems where these cyber-physical systems are formed ad hoc according to the specific local topology, the available networking capabilities, and the goals and constraints of the subsystems captured by the information processing part. In this paper we present a formalism that permits to model the sketched class of cyber-physical systems. The ad hoc formation of tightly coupled subsystems of arbitrary size are specified using a UML-based graph transformation system approach. Differential equations are employed to define the resulting tightly coupled behavior. Together, both form hybrid graph transformation systems where the graph transformation rules define the discrete steps where the topology or modes may change, while the differential equations capture the continuous behavior in between such discrete changes. In addition, we demonstrate that automated analysis techniques known for timed graph transformation systems for inductive invariants can be extended to also cover the hybrid case for an expressive case of hybrid models where the formed tightly coupled subsystems are restricted to smaller local networks.
Service-oriented modeling employs collaborations to capture the coordination of multiple roles in form of service contracts. In case of dynamic collaborations the roles may join and leave the collaboration at runtime and therefore complex structural dynamics can result, which makes it very hard to ensure their correct and safe operation. We present in this paper our approach for modeling and verifying such dynamic collaborations. Modeling is supported using a well-defined subset of UML class diagrams, behavioral rules for the structural dynamics, and UML state machines for the role behavior. To be also able to verify the resulting service-oriented systems, we extended our former results for the automated verification of systems with structural dynamics [7, 8] and developed a compositional reasoning scheme, which enables the reuse of verification results. We outline our approach using the example of autonomous vehicles that use such dynamic collaborations via ad-hoc networking to coordinate and optimize their joint behavior.
Creating fonts is a complex task that requires expert knowledge in a variety of domains. Often, this knowledge is not held by a single person, but spread across a number of domain experts. A central concept needed for designing fonts is the glyph, an elemental symbol representing a readable character. Required domains include designing glyph shapes, engineering rules to combine glyphs for complex scripts and checking legibility. This process is most often iterative and requires communication in all directions. This report outlines a platform that aims to enhance the means of communication, describes our prototyping process, discusses complex font rendering and editing in a live environment and an approach to generate code based on a user’s live-edits.
Industrial policy and social strategy at the corporate level in Poland : questionnaire results
(1999)
This paper presents results from a survey of industrial policy of the state and the social security system at the corporate level in Poland. Previous reports in this area indicated preferable directions of research to be taken in order to prove various hypotheses of the purposefulness of an integral approach to industrial policy and social security in the analysis of economic processes in transition (see Weikard 1997). This paper summarises the results and draws conclusions from a questionnaire study on subsidies, social benefits and economic policy in Polish firms during the process of transformation. Our results and conclusions show the scope and character of the processes in the area of industrial and social policy in the period 1994 to 1997. The paper is divided into five parts. The first part concerns the aims and methodology of the questionnaire; it also gives a brief description of the sample. The second part shows how enterprises dealt with the issues of employment and wages in this period. The third part characterises industrial policy at the corporate level, while the next presents results from the survey of various social schemes pursued. The final part aims at an integral approach in the analysis of various processes taking place in Polish enterprises. The survey was conducted in the period April to June 1998. Its aim was to observe certain phenomena occurring at the corporate level. The questionnaire was distributed among the managers, directors and presidents of large-size enterprises, which had been selected to satisfy the following three criteria. Firstly, the number of employees had to be considerable (over 300 workers). This criterion was applied following the consideration that certain social phenomena are more conspicuous in enterprises with large manpower. Secondly, only operating enterprises were selected, the enterprises which closed down were disregarded. Finally, for the purposes of the survey the units differed as regards their legal situation and form of ownership. Out of over 1800 enterprises 370 units were drawn where we sent the questionnaire. Unfortunately, as many as 51.9% of the respondents refused co-operation, questions to a certain extent puts the representativeness of the sample in question. Finally, 178 questionnaires were subsequently completed and returned for analysis. However, not all of these questionnaires included full answers to all of the 75 questions; therefore, while discussing the results of the survey we have indicated the number of relevant answers we have received.
Legal protection against breaches of duty on the part of the German works council : a fata morgana?
(2000)
Graph queries have lately gained increased interest due to application areas such as social networks, biological networks, or model queries. For the relational database case the relational algebra and generalized discrimination networks have been studied to find appropriate decompositions into subqueries and ordering of these subqueries for query evaluation or incremental updates of query results. For graph database queries however there is no formal underpinning yet that allows us to find such suitable operationalizations. Consequently, we suggest a simple operational concept for the decomposition of arbitrary complex queries into simpler subqueries and the ordering of these subqueries in form of generalized discrimination networks for graph queries inspired by the relational case. The approach employs graph transformation rules for the nodes of the network and thus we can employ the underlying theory. We further show that the proposed generalized discrimination networks have the same expressive power as nested graph conditions.
Graph databases provide a natural way of storing and querying graph data. In contrast to relational databases, queries over graph databases enable to refer directly to the graph structure of such graph data. For example, graph pattern matching can be employed to formulate queries over graph data.
However, as for relational databases running complex queries can be very time-consuming and ruin the interactivity with the database. One possible approach to deal with this performance issue is to employ database views that consist of pre-computed answers to common and often stated queries. But to ensure that database views yield consistent query results in comparison with the data from which they are derived, these database views must be updated before queries make use of these database views. Such a maintenance of database views must be performed efficiently, otherwise the effort to create and maintain views may not pay off in comparison to processing the queries directly on the data from which the database views are derived.
At the time of writing, graph databases do not support database views and are limited to graph indexes that index nodes and edges of the graph data for fast query evaluation, but do not enable to maintain pre-computed answers of complex queries over graph data. Moreover, the maintenance of database views in graph databases becomes even more challenging when negation and recursion have to be supported as in deductive relational databases.
In this technical report, we present an approach for the efficient and scalable incremental graph view maintenance for deductive graph databases. The main concept of our approach is a generalized discrimination network that enables to model nested graph conditions including negative application conditions and recursion, which specify the content of graph views derived from graph data stored by graph databases. The discrimination network enables to automatically derive generic maintenance rules using graph transformations for maintaining graph views in case the graph data from which the graph views are derived change. We evaluate our approach in terms of a case study using multiple data sets derived from open source projects.
Die vorliegende Arbeit stellt eine kritische Übersicht über den Forschungsstand zu multiplen Wh-Konstruktionen im Slavischen dar. Das Ziel ist es, die Unklarheit der Datenlage und die Widersprüchlichkeit der auf solchen "unklaren" Daten basierten Theorien aufzuzeigen. Inhalt: Historischer Hintergrund (Wachowicz 1974) Einige ältere Ansätze Höhepunkt: die folgenschwere Arbeit von Rudin (1988) Probleme: - Das Problem der Zuverlässlichkeit von Daten - Das Problem der Relevanz von Daten "Harte" Fakten: - Strikte Superioritätseffekte im Bulgarischen - Obligatorische Wh-Anhebung im Slavischen Neuere Ansätze: - "Qualitative" Ansätze - "Quantitative" Ansätze - Alternative Ansätze
TripleA is a workshop series founded by linguists from the University of Tübingen and the University of Potsdam. Its aim is to provide a forum for semanticists doing fieldwork on understudied languages, and its focus is on languages from Africa, Asia, Australia and Oceania. The second TripleA workshop was held at the University of Potsdam, June 3-5, 2015.
The route to chaos for a two-dimensional externally driven flow : [to appear in Physical Review E]
(1998)
The conidial stage and chasmothecia of Golovinomyces orontii have been found in Germany on cultivated Limnanthes douglasii. A powdery mildew anamorph found in the Netherlands on Malva alcea agrees morphologically with the Oidium of the latter species as well. Golovinomyces sp. (anamorph) on Parthenium integrifolium is described and discussed. Erysiphe sp. has been found in Germany on Acer opalus, and E. magnifica is recorded from Germany and Switzerland on Magnolia spp. Oidium passiflorae is new to Switzerland. An Oidium morphologically agreeing with the anamorph of Podosphaera aphanis has recently been collected on Exacum macranthum cultivated in a greenhouse, and conidiophores and conidia of a species of Podosphaera sect. Sphaerotheca subsect. Magnicellulatae (P. fusca complex) on Phlox paniculata and Polemonium caeruleum have been found in Germany.
The introduction of a new powdery mildew disease on Rhus hirta in various parts of Germany (Brandenburg, Rhine- Westphalia, Sachsen-Anhalt and Saxony) is reported. The anamorph found on this host agrees well with the North American Podosphaera pruinosa. Although the teleomorph has not yet been found in Germany and a molecular study has not yet been possible due to the lack of fresh North American material for a comparison, there is little doubt that the European outbreak of the Rhus powdery mildew disease may be referred to as Podosphaera pruinosa. Morphology, taxonomy and distribution of Podosphaera species on Rhus and other hosts of the Anacardiaceae are discussed in detail.
Since 2002, keywords like service-oriented engineering, service-oriented computing, and service-oriented architecture have been widely used in research, education, and enterprises. These and related terms are often misunderstood or used incorrectly. To correct these misunderstandings, a deeper knowledge of the concepts, the historical backgrounds, and an overview of service-oriented architectures is demanded and given in this paper.
The intensive development of industry and urban structures along the seashores of the world, as well as the immense increase in marine transportation and other activities, has resulted in the deposition of thousands of new chemicals and organic compounds, endangering the existence of organisms and ecosystems. The conventional single biomarker methods used in ecological assessment studies cannot provide an adequate base for environmental health assessment, management and sustainability planning. The present study uses a set of novel biochemical, physiological, cytogenetic and morphological methods to characterize the state of health of selected molluscs and fish along the shores of the German North Sea, as well as the Israeli Mediterranean and Red Sea. The methods include measurement of activity of multixenobiotic resistance-mediated transporter (MXRtr) and the system of active transport of organic anions (SATOA) as indicators of antixenobiotic defence; glutathione S-transferase (GST) activity as an indicator of biotransformation of xenobiotics; DNA unwinding as a marker of genotoxicity; micronucleus test for clastogenicity; levels of phagocytosis for immunotoxicity; cholinesterase (ChE) activity and level of catecholamines as indicators of neurotoxicity; permeability of external epithelia to anionic hydrophilic probe, intralysosomal accumulation of cationic amphiphilic probe and activity of non-specific esterases as indicators of cell/tissue viability. Complete histopathological examination was used for diagnostics of environmental pathology. The obtained data show that the activity of the defensive pumps, MXRtr and SATOA in the studied organisms was significantly higher in the surface epithelia of molluscs from a polluted site than that of the same species from control, unpolluted stations, providing clear evidence of response to stress. Enhanced frequency of DNA lesions (alkaline and acidic DNA unwinding) and micronucleus-containing cells was significantly higher in samples from polluted sites in comparison to those from the clean sites that exhibited genotoxic and clastogenic activity of the pollutants. In all the studied molluscs a negative correlation was found between the MXRtr levels of activity and the frequency of micronucleus-containing hemocytes. The expression of this was in accordance with the level of pollution. The complete histopathological examination demonstrates significantly higher frequencies of pathological alterations in organs of animals from polluted sites. A strong negative correlation was found between the frequency of these alterations and MXRtr activity in the same specimens. In addition to these parameters, a decrease in the viability was noted in molluscs from the polluted sites, but ChE activities remained similar at most sites. The methods applied in our study unmasked numerous early cryptic responses and negative alterations of health in populations of marine biota sampled from the polluted sites. This demonstrates that genotoxic, clastogenic and pathogenic xenobiotics are present and act in the studied sites and this knowledge can provide a reliable base for consideration for sustainable development