- informatik-vortraege - lists.rwth-aachen.de

Oberseminar 28.11.
by Blundell, Inga 29 Oct '18

29 Oct '18

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Mittwoch, 28. November 2018, 10.00 Uhr Ort: E1, Raum 4312, Ahornstr. 55 Referentin: Inga Blundell Thema: Algorithmic solver selection for systems of differential equations in human brain cell models for the domain specific language NESTML Abstract: The fast advancements in neuroscience and the multitude of newly developed neuron models require a language to describe neurons by means of systems of differential equations. The capability of implementing new models in precise agreement with their mathematical definition and subsequently running computationally efficient simulations is essential for progress on the theoretical foundation of neuroscience. The use of abstract modelling languages like NESTML enables neuroscientists to formally specify their models and to automatically produce high level simulation code that runs efficiently on various computer architectures. An important requirement to these modelling languages is the automatic generation of efficient and precise integrations schemes for differential equations that describe the dynamical behaviour of neuron models. An algorithmic approach to this requirement will be presented in this talk. Es laden ein: Die Dozenten der Informatik ------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------ Forschungszentrum Juelich GmbH 52425 Juelich Sitz der Gesellschaft: Juelich Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498 Vorsitzender des Aufsichtsrats: MinDir Dr. Karl Eugen Huthmacher Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender), Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt, Prof. Dr. Sebastian M. Schmidt ------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------

1 0

Informatik-Kolloquium
by Fachgruppensprecher 24 Oct '18

24 Oct '18

This Friday Zeynep Akata will give a talk about “Explaining and Representing Novel Concepts With Minimal Supervision”. You are kindly invited. Please see the details below. Best regards, Martin Grohe ————— Title: Explaining and Representing Novel Concepts With Minimal Supervision Speaker: Zeynep Akata (University of Amsterdam) Time and Place: Friday, October 26th, 14:15, AH I Abstract: Clearly explaining a rationale for a classification decision to an end-user can be as important as the decision itself. Existing approaches for deep visual recognition are generally opaque and do not output any justification text; contemporary vision-language models can describe image content but fail to take into account class-discriminative image properties which justify visual predictions. In this talk, I will present my past and current work on Zero-Shot Learning, Vision and Language for Generative Modeling and Explainable Artificial Intelligence where we show (1) how to generalize image classification models to cases when no visual training data is available, (2) how to generate images and image features using detailed visual descriptions, and (3) how our models focus on discriminating properties of the visible object, jointly predict a class label, explain why/not the predicted label is chosen for the image. ---------------------------------------------------------------- Prof. Dr. Martin Grohe | Fachgruppe Informatik http://www.informatik.rwth-aachen.de/ | Ahornstr. 55, 52074 Aachen, Germany | RWTH Aachen University Phone ++49 241 80-21701 / Fax -22215 |

1 0

Einladung Informatik-Oberseminar
by Matthias Terber 22 Oct '18

22 Oct '18

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Montag, 29. Oktober 2018, 10:00 Uhr Ort: Raum 2202, i11, Ahornstr. 55 Referent: Matthias Terber, M. Sc. Thema: Praxisorientierte Verwendung und Evaluierung synchroner Programmierung in reaktiven, eingebetteten Systemen Abstract: Durch die große Dominanz der Programmiersprache C im industriellen Bereich, werden die Probleme von reaktiven, eingebetteten Systemen üblicherweise mit konventioneller, sequentieller Programmierung adressiert. Diese bietet jedoch keine spezifische Unterstützung für die Beschreibung von reaktivem Verhalten, wodurch Entwicklung, Verständlichkeit und Wartbarkeit der Software erschwert werden. Synchrone Programmiersprachen könnten eine vielversprechende Lösung darstellen, um den Entstehungsprozess von Software zu vereinfachen und deren Qualität zu verbessern. Bis heute ist der Einsatz von synchroner Programmierung jedoch nur auf sehr spezielle industrielle Anwendungsfälle beschränkt; ihre praktische Anwendung und deren Nutzen werden in existierender Literatur kaum behandelt. In dieser Arbeit wird eine Fallstudie durchgeführt, welche die Anwendbarkeit und Eignung des synchronen Programmierparadigmas anhand eines konkreten, existierenden Smart Devices aus der Industrie untersucht. Mit Fokus auf die reaktive Domäne werden die technischen Herausforderungen für die Entwicklung sowie die Qualitätsprobleme der bestehenden Softwarelösung herausgearbeitet. Basierend auf der synchronen Sprache Céu wird eine synchrone Neuimplementierung skizziert. Diese veranschaulicht den Einsatz von synchroner Programmierung und zeigt deren Vorteile für die Softwareentwicklung. Schließlich glauben wir, dass diese Arbeit grundsätzlich einen praktikablen Weg aufzeigt, wie sich die Qualität eingebetteter Software in industriellen Anwendungen verbessern lässt. Es laden ein: Die Dozenten der Informatik

1 0

Einladung: Oberseminar Herr Tim Lange
by Tim Lange 27 Sep '18

27 Sep '18

************************************************************************ ** * Einladung * * * Informatik-Oberseminar * * ************************************************************************ Zeit: Freitag, 5. Oktober 2018, 10:00 Uhr Ort: Gebäude E2, B-IT-Seminarraum 5053.1, Ahornstr. 55 Referent: Herr Tim Lange, M.Sc. Titel: IC3 Software Model Checking Abstract: Many real-world systems are becoming ever smaller in size and more powerful, such that software becomes more complex. The risk of ubiquitous softwares' misbehaviour and the resulting damage therefore grows dramatically. In order to prevent such erroneous behaviour, model checking, a formal verification technique for determining functional properties of information and communication systems, has proven to be highly useful. For proving mathematical properties, one of the first methods to be taught in schools is induction. With the concept of proving a concrete induction base and a general induction step it is considered a very simple and intuitive, yet powerful proof method. However, finding an inductive formulation for difficult properties can be an extremely hard task. When humans try to solve this problem, they naturally produce a set of smaller, simple lemmas that together imply the desired property. Each of these lemmas holds relative to some subset of previously established lemmas by invoking the knowledge to prove the new lemma. This incremental approach to proving complex properties using sets of small inductive lemmas was first applied to model checking of hardware systems in the IC3 algorithm and has proven to outperform all known approaches to hardware model checking. This talk aims at applying the principles of incremental, inductive verification laid by the IC3 algorithm to software model checking for industrial control software with special attention to the control-flow induced by the program under consideration. For this purpose, basic concepts are introduced and an in-depth explanation of the IC3 algorithm and its different building blocks (search phase, generalization and propagation) is given. Based on these prerequisites, the novel IC3CFA algorithm is presented. In this algorithm, the control-flow of the program is explicitly modelled as an automaton, while the variable valuations are handled symbolically, thus using the best of both worlds. Following the search phase of IC3CFA, solutions for applying generalization to IC3CFA are presented. Finally, the performance of the IC3CFA algorithm and all proposed improvements is extensively evaluated on a set of well-recognised benchmarks. Es laden ein: Die Dozenten der Informatik

1 0

Talk by Juergen Gall, Wed 19.09. 15h in UMIC 025
by Bastian Leibe 18 Sep '18

18 Sep '18

Hi everybody, We're happy to announce the following talk by Prof. Juergen Gall (University of Bonn), which will take place tomorrow, Wednesday, at 15:00h in room UMIC 025. Anybody interested is welcome to attend. Please feel free to forward the announcement. Best regards, Bastian Leibe Speaker: Prof. Juergen Gall (Uni Bonn) Time: Wed. 19.09.2018, 15:00h Place: UMIC 025 Title: Analyzing and Anticipating Human Behavior in Videos -- _______________________________________________________________ Prof. Dr. Bastian Leibe Informatik 8 (Computer Vision) UMIC Research Centre RWTH Aachen University Mies-van-der-Rohe-Strasse 15 office: 124 DE-52074 Aachen phone: +49-241-80 20762 Germany mailto:leibe@vision.rwth-aachen.de _______________________________________________________________

1 0

Einladung: Informatik-Oberseminar Sebastian Freitag
by Freitag, Sebastian 13 Sep '18

13 Sep '18

*********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * *********************************************************************** Zeit: Donnerstag, 27. September 2018, 09:30 Uhr Ort: Seminarraum 003 des IT-Centers, Kopernikusstr. 6 Referent: Sebastian Freitag, M. Sc. Titel: Supported Navigation in Immersive Virtual Environments Abstract: Navigation - the process of getting from one location in a virtual scene to another - is one of the most important and fundamental types of interaction in Virtual Reality, as it is part of most use cases and applications in some way or another. However, there is no single navigation technique that is suitable and appropriate for all use cases, as there is typically a different set of requirements for each application, and all techniques come with their own set of drawbacks. Therefore, the selection of a navigation technique for any use case is usually a trade-off decision. Therefore, this work investigates and proposes navigation support techniques to mitigate the negative aspects of these trade-offs. First, ways to manipulate a user's movement to increase the share of real walking relative to the overall amount of navigation are examined. In this context, a navigation technique is presented that repositions the user while they travel in order to maximize the interaction space for real walking at the target location. Furthermore, it is investigated whether rotation gain (the manipulation of the mapping between physical and virtual head rotations), which can be used for techniques such as redirected walking, is usable in screen-based setups such as CAVE-like environments. Second, the work proposes navigation support techniques based on an automated analysis of the virtual scene, that are aimed at making navigation more efficient. To this end, we analyze scene visibility - which parts of the scene are visible from any viewpoint - and the evaluation of the quality of such a viewpoint. In this context, the work compares different measures to estimate viewpoint quality, contributing several novel methods, and introduces an approach for approximating scene visibility efficiently by restricting the computation to navigable areas of a scene. It is further shown that these techniques can be used successfully to increase the efficiency of navigation, proposing a technique to adjust travel speed automatically, reducing cognitive load, an approach to find intermediate target locations during long-distance travel to reduce cybersickness, and an assistance technique that improves cognitive map buildup during exploration. Es laden ein: Die Dozenten der Informatik

1 0

Einladung: Informatik-Oberseminar Harold Bruintjes
by Harold Bruintjes 07 Sep '18

07 Sep '18

************************************************************************ * * * Einladung * * * Informatik-Oberseminar * * ************************************************************************ Zeit: Donnerstag, 27. September 2018, 10.30 Uhr Ort: E1, Raum 5053.2, Ahornstr. 55 Referent: Harold Bruintjes, M.Sc. Topic: Model-Based Reliability Analysis of Aerospace Systems Abstract: In order to provide reliable and safe systems in the aerospace domain, despite increased complexity and stronger demands on capabilities, new model-based system and software engineering approaches have to be developed. The COMPASS project, an international research project involving RWTH, ESA, FBK and various industrial partners, delivers a toolset for precisely this purpose. In this talk, the approach taken by the COMPASS toolset is discussed, in particular with regard to its most recent advancements. After an introduction, the talk will focus on new results in requirements formalization, the use of statistical model checking, the capabilities of the toolset itself and its application to a case study. COMPASS offers a multitude of analyses pertaining to correctness, safety, reliability and fault management, backed by a single modeling formalism derived from AADL. More information about COMPASS can be found on its website: http://www.compass-toolset.org . Es laden ein: Die Dozenten der Informatik

1 0

Einladung: Informatik-Oberseminar Ana-Maria-Cristina Nicolaescu
by Marion Zinner 28 Aug '18

28 Aug '18

************************************************************************ * * * Einladung * * * * Informatik-Oberseminar * * * ************************************************************************ Zeit: Donnerstag, 13 September 2018, 10:00 Ort: E2, Raum 5056, Ahornstraße 55 Referent: Ana-Maria-Cristina Nicolaescu, M.Sc. Titel: Behavior-Based Architecture Conformance Checking Abstract: In the past years several approaches for static-based architecture conformance checking were proposed. These pose important limitations when considering modern systems, typically composed of several interacting processes. Ever since the advent of object orientation and due to the shift from monolith architectures to componentized ones, the complexity of software systems has moved from structure to behavior. This is typically out of the scope of static-based conformance approaches, which face an impossibility in assessing if the system under analysis is behaving as foreseen by its architects. ARAMIS is our solution to alleviating the above-mentioned problem. First, the intended architecture description of the system under analysis is expressed using an ARAMIS-specific meta-model. This encompasses the architecture units constituting the system and the communication rules governing these. To increase acceptance, model-engineering techniques are also proposed to enable the reuse of intended architecture descriptions elaborated using different meta-models than that of ARAMIS. Next, interactions are extracted during the system's execution using third party monitoring tools. Given that a holistic analysis of the behavior of a system is impossible in general, we proposed several indicators to assess whether the captured interactions represent an adequate basis for checking the conformance of the system as a whole. The interactions are consequently elevated to depict communication between the units defined in the system's intended architecture description and validated to check their conformance to the formulated communication rules. The results constitute a description of the implemented architecture of the system, characterized by its drift from the intended one. This can subsequently be explored using several mechanisms such as user-defined architecture views and perspectives or dedicated visualizations. Last but not least, processes for guiding the activities involved in behavior-based conformance checking were developed and described. The ARAMIS concept and the developed toolbox were evaluated in three case studies, the last two being conducted in industrial settings. The results were very positive. The evaluation proved that the ARAMIS approach can be utilized by organizations when attempting to understand and evaluate the current state of implemented architectures and as a starting point for future evolution. Es laden ein: Die Dozenten der Informatik

1 0

Einladung Informatik-Oberseminar
by Hermerschmidt, Lars 17 Aug '18

17 Aug '18

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Montag, 29. Oktober 2018, 15:00 Uhr Ort: Raum 9U10, Gebäude E3, Ahornstr. 55 Referent: Dipl. Inform. Lars Hermerschmidt Titel: Agile Modellgetriebene Entwicklung von Software Security & Privacy Abstract: IT systems continue to penetrate all aspects of daily life, so that these systems are required to function correctly while they aggregate more and more information to enable a more comfortable and efficient life. However, every day 16 new vulnerabilities in average are reported in software products, which enable adversaries to abuse system functionality in developer unintended way. Nevertheless users do not only fear unauthorized access to their data by adversarial third party when they compromise system's security, but as well if services do not provide transparency over their collection and processing of sensitive user data. These security and privacy aspects need to be addressed during software system development, where developers typically focus on a constructive view on the system, which makes it hard for them to think like an adversary who misuses the system. In this thesis the three aspects security architecture, correct handling of input data during the creation of output data, and user privacy are researched in the context of software development. Models are used to describe these aspects and the agile model-driven software development approach is utilized to connect those models to executable systems. To enable developers to express their constructive view on the security architecture of a system, the security architecture modeling language MontiSecArc is presented. From this language a domain specific transformation language is derived, which enables security experts to express known flaws in the security architecture together with a solution in a flaw correction pattern. These pattern can be applied automatically within an agile development process to prevent architectural flaws. In addition the notation of flaw correction patterns enables precise description, naming, and distinction of architectural flaws such that flaw correction patterns can be collected. Injection vulnerabilities like Cross-Site-Scripting (XSS) and SQL Injection are the most common class of vulnerabilities, which emerge if a program uses input data to create output data that is written in a language without encoding the input according to the output's language. This injection vulnerability preventing encoding is depending on the output language and the different contexts within the language. Therefore, in this work the MontiCoder approach is presented, which integrates the context-specific encoding into the language definition and derives unparser and parser from this definition, which automatically perform context-specific encoding and decoding. This approach shifts the definition of the correct encoding of data from the developer who is using a language to the language developer and provides the latter one the ability to precisely define the encoding together with the language itself. Facing the challenges of developers who aim to provide more transparency and self-determinism to service users about the usage of there data within a service, in this thesis the Privacy Development Language (PDL) is presented. The PDL enables developers to model the service's data structure along with the usage of the data within the service. From this data model an interactive privacy policy is generated, which enables users to select those parts of the service they want to use and present them a description of the used data and the procession which is performed to deliver the selected service. This way the approach enables service users to perform an informed privacy decision and eases developers work to provide transparency. Explicitly modeling these aspects enables analyses of security an privacy on model level, such that vulnerabilities are fixed within models as well. This way there is no need to extract all flaws from handwritten low level code after development. Es laden ein: Die Dozenten der Informatik

1 0

Einladung: Informatik-Oberseminar
by Hölldobler, Katrin 16 Aug '18

16 Aug '18

+********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +********************************************************************** Zeit: Dienstag, 11. September 2018, 10.00 Uhr Ort: E3, Raum 9222 Referent: Katrin Hölldobler, M.Sc. RWTH Thema: MontiTrans: Agile, modellgetriebene Entwicklung von und mit domänenspezifischen, kompositionalen Transformationssprachen Abstract: Modelle sind die zentralen Entwicklungsartefakte der modellgetriebenen Softwareentwicklung und müssen entsprechend überarbeitet, weiterentwickelt und gewartet werden. Daher sind Modelltransformationen ebenfalls essenziell für die modellgetriebene Softwareentwicklung. Während domänenspezifische Sprachen (DSLs) zur Modellierung mittlerweile weitverbreitet sind, sind spezifische Transformationssprachen rar. Stattdessen werden General Purpose Transformationssprachen (GPTLs) verwendet, die Transformationen basierend auf der internen Repräsentation der Modelle formulieren. Diese interne Repräsentation ist Modellierern und Domänenexperten in der Regel unbekannt, was deren Einbindung in den Entwicklungsprozess deutlich erschwert. Domänenspezifische Transformationssprachen (DSTLs) basieren auf der den Modellierern bekannten konkreten Syntax der DSL, wodurch sie spezifisch für die zugehörige DSL sind. Dies verringert den initialen Aufwand zum Erlernen der Transformationssprache, da der größte Teil der Syntax bereits bekannt ist und zusätzlich nur die Syntax der Transformationsoperatoren erlernt werden muss. Andererseits haben DSTLs durch ihre Zugehörigkeit zu einer DSL den Nachteil, dass für jede neu entwickelte DSL gleichzeitig oder für existierende DSLs nachträglich eine DSTL entwickelt werden muss. Aus diesem Grund erhöht sich der Aufwand der DSL-Entwicklung deutlich. Darüber hinaus erhöht sich der Aufwand weiter, wenn DSTLs zur Übersetzung zwischen verschiedenen Sprachen benötigt werden und entwickelt werden müssen. Zur Reduzierung dieses Aufwands wurde in [Wei12] ein erster DSTL-Generator vorgestellt. Dieser wurde mithilfe der Language Workbench MontiCore 2 entwickelt. Besondere Merkmale von MontiCore 4 sind die Möglichkeiten zur modularen Sprachdefinition durch Sprachkomposition sowie die an die Objektorientierung angelehnten Konzepte der Nichtterminalerweiterung als Erweiterung des Alternativenkonzepts von Grammatiken. Der bisherige DSTL-Generator ermöglicht eine Generierung von DSTLs aus MontiCore-Grammatiken und unterstützt ausschließlich monolithische Sprachdefinitionen. Die Mehrzahl der mit MontiCore entwickelten Modellierungssprachen sind hingegen kompositional definiert. Die bisher generierten DSTLs unterstützen ausschließlich Transformationen von Modellen innerhalb einer Modellierungssprache. Übersetzungen zwischen verschiedenen Sprachen oder die Migration von Modellen zwischen Sprachversionen ist somit nicht möglich. Des Weiteren wird bisher nur ein Teil der möglichen Konzepte von MontiCore-Grammatiken unterstützt, auf den sich die restlichen Konzepte abbilden lassen. Dadurch sind Sprachentwickler in der Sprachdefinition eingeschränkt. Im Rahmen dieser Dissertation wird die generative Entwicklung und Verwendung von DSTLs in der modellgetriebenen Softwareentwicklung durch MontiTrans unterstützt. MontiTrans ermöglicht die Entwicklung neuer DSTLs und der zugehörigen Infrastruktur zur Spezifikation und Ausführung von Modelltransformationen. Für die Entwicklung von MontiTrans wurden die zuvor beschriebenen offenen Punkte aufgegriffen und die Generierung von DSTLs basierend auf den Ergebnissen aus [Wei12] weiterentwickelt. MontiTrans wurde sowohl zur Entwicklung neuer DSTLs als auch zur Entwicklung von Bibliotheken von wiederverwendbaren Transformationen verwendet. Hierdurch konnte gezeigt werden, dass MontiTrans ein umfassendes Werkzeug zur Entwicklung von DSTLs sowie für die Entwicklung von Modelltransformationen innerhalb modellgetriebener Softwareentwicklungsprojekte ist. MontiTrans erleichtert sowohl Sprachentwicklern die Entwicklung neuer DSLs und zugehöriger DSTLs als auch Transformationsentwicklern die Definition und Anwendung neuer Transformationen. Es laden ein: Die Dozenten der Informatik

1 0