@techreport{Schmitt2024, author = {Schmitt, Ingo}, title = {ANN-Partitionen und CQQL-Ausdr{\"u}cke}, doi = {10.26127/BTUOpen-6606}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-66069}, year = {2024}, abstract = {Gegeben sei f{\"u}r ein bin{\"a}res Klassifikationsproblem ein k{\"u}nstliches, neuronales Netzwerk ann bestehend aus ReLUKnoten und linearen Schichten (convolution, pooling, fully connected). Das Netzwerk ann sei mit hinreichender Genauigkeit an Hand von Trainingsdaten trainiert. Wir werden zeigen, dass ein solches Netzwerk in verschiedene Partitionen des Eingaberaums zerlegt werden kann, wobei jede Partition eine lineare Abbildung der Eingabewerte auf einen klassifizierenden Ausgabewert repr{\"a}sentiert. Im Weiteren gehen wir von einem einfachen Netzwerk ann aus, bei dem die Eingangswerte Mintermen von Attributwerten entsprechen. Einfach ist ein Netzwerk, wenn es f{\"u}r eine geringe Anzahl von Attributen trainiert wurde und die Anzahl der ReLU-Knoten ebenfalls gering ist. In der Arbeit wird gezeigt, dass jede lineare Partition durch einen CQQL-Ausdruck beschrieben werden kann. Ein CQQL-Ausdruck l{\"a}sst sich mit Hilfe von Quantenlogik-inspirierten Entscheidungsb{\"a}umen beschreiben.}, subject = {Quantum logic; Artificial neural networks; Quantenlogik; K{\"u}nstliche neuronale Netze; Neuronales Netz; Quantenlogik}, language = {de} } @techreport{Schmitt2021, author = {Schmitt, Ingo}, title = {Generating CQQL conditions from classifying CNNs}, doi = {10.26127/BTUOpen-5550}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-55503}, year = {2021}, abstract = {Convolutional neural networks are often successfully used for classification problems. Usually, a huge number of weights need to be learnt by use of training data. However, the learnt weights give no insight how the cnn really works. Thus, a cnn can be seen as a black box solution. In our approach we develop a method to generate a commuting quantum query language (cqql) condition from a sample derived from a given cnn or from training input. CQQL is inspired by quantum logic and its conditions obey the rules of Boolean algebra. The evaluation of a cqql condition provides values from the unit interval [0; 1] and establishes therefore an elegant bridge between logic and a cnn. The underlying assumption is that a condition (a logic expression) gives much more understanding than pure cnn weights. Furthermore, the rich theory of Boolean algebra can be used for manipulating logic expressions. After extracting a cqql condition from a cnn or its training data we can use logic as a way to predict classes alternatively to a cnn.}, subject = {CQQL; Quantum Logic; CNN; CQQL; Quantenlogik; CNN; Erkl{\"a}rbarkeit; Zellulares neuronales Netzwerk; Abfragesprache; Quantenlogik}, language = {en} } @techreport{Schmitt2019, author = {Schmitt, Ingo}, title = {Quantum-based construction of a probability measure}, editor = {Schmitt, Ingo}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-50760}, year = {2019}, abstract = {From Gleason's theorem we know that in principle every probability measure can be expressed by Hermitian operators in a separable Hilbert space and the Born rule as part of a quantum mechanical system. However, that theorem is not constructive. For a given discrete and additive probability measure based on a σ-algebra we construct a quantum system with projectors expressing that probability measure.}, subject = {Quantum modelling; Probability measure; Wahrscheinlichkeitsmetrik; Mathematische Modellierung; Quanteninformatik}, language = {en} } @techreport{HerajyLiuRohretal.2017, author = {Herajy, Mostafa and Liu, Fei and Rohr, Christian and Heiner, Monika}, title = {(Coloured) Hybrid Petri nets in Snoopy - user manual}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-41572}, year = {2017}, abstract = {Hybrid simulation of biological processes becomes widely used to overcome the limitations of the pure stochastic or the complete deterministic simulation. In this manual, we present easy-to-follow steps for constructing and executing hybrid models via Snoopy [HHL+12]. Snoopy is a tool to design and animate or simulate hierarchical graphs, i.e., qualitative, stochastic, continuous, and hybrid Petri nets. This manual is concerned with hybrid Petri nets (HPN) [HH12] as well as their coloured counterpart (HPNC) [HLR14]. HPN combine the merits of stochastic and continuous Petri nets into one single class. Moreover, HPN in Snoopy supports state of the art hybrid simulation algorithms (e.g., [HH16]) to execute the constructed HPN models. Simulating a model using Snoopy's hybrid simulation involves first constructing the reaction network via HPN notations and afterwards executing such model.}, subject = {Petri net; Hierarchical graph; Gef{\"a}rbtes Petri-Netz; Hybridsimulation}, language = {en} } @techreport{SchwarickRohrHeiner2016, author = {Schwarick, Martin and Rohr, Christian and Heiner, Monika}, title = {MARCIE manual}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-40568}, year = {2016}, abstract = {This manual gives an overview on MARCIE - Model Checking And Reachability analysis done effiCIEntly. MARCIE was originally developed as a symbolic model checker for stochastic Petri nets, building on its predecessor - IDDMC - Interval Decision Diagram based Model Checking - which has been previously developed for the qualitative analysis of bounded Place/Transition nets extended by special arcs. Over the last years the tool has been enriched to allow also quantitative analysis of extended stochastic Petri nets. We concentrate here on the user viewpoint. For a detailed introduction to the relevant formalisms, formal definitions and algorithms we refer to related literature.}, subject = {Petri nets; Model checking; Model Checking; Petri-Netz}, language = {en} } @techreport{LieskeNoackSchwalbeetal.2016, author = {Lieske, Henry and Noack, Daniel and Schwalbe, Christoph and Vogel, Elisabeth and Hinze, Thomas}, title = {Faszination Molekulares Rechnen: Studentische Beitr{\"a}ge vermitteln Erkenntnisse interdisziplin{\"a}rer Forschung}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-40328}, year = {2016}, abstract = {Molekulares Rechnen etabliert sich als interdisziplin{\"a}r gepr{\"a}gtes Wissensgebiet mit hohem Anwendungspotential. Es besch{\"a}ftigt sich mit Prinzipien biologischer Informationsverarbeitung. Molek{\"u}le bzw. molekulare Systeme {\"u}bernehmen die Rolle eines Datentr{\"a}gers, auf dem biochemische Prozesse oder physische Wechselwirkungen operieren. Facetten sind unter anderem DNA-Computing, chemisches Rechnen, Membrane Computing und ameisenbasiertes Computing. Vier studentische Beitr{\"a}ge stellen diese Forschungsrichtungen anhand eing{\"a}ngiger Beispiele anschaulich vor.}, subject = {Molekulares Rechnen; DNA Computing; Chemisches Rechnen; Membrane Computing; Ameisenbasiertes Computing; DNA computing; Chemical computing; Membrane computing; Molecular computing; Ant-based computing; Biocomputer; Molekulare Bioinformatik}, language = {de} } @techreport{Sinha2014, author = {Sinha, Aman}, title = {Comparison of stochastic simulation tools}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-31460}, year = {2014}, abstract = {This report compares some stochastic simulation tools for biochemical reaction networks. The stochastic simulation tools are selected on the basis of some selection criteria. Simulations are performed for the different stochastic simulation tools on different benchmark models. This report gives an overview of how the comparison is carried out for the chosen tools. The tools are compared on a common evaluation protocol. The evaluation protocol comprises a set of benchmark models along with the parameters which are provided as input to the tools. The benchmark models are represented as Petri nets and fed in SBML (System Biology Markup Language) to the different tools. Experiments are performed on each tool and the results are recorded. The tools are finally compared based on the comparison criteria.}, subject = {Benchmarks; Biochemical reaction networks; Evaluation protocol; SBML; Stochastic simulation; Stochastisches Modell; Petri-Netz}, language = {en} } @techreport{Swapnil2014, author = {Swapnil, Chiru}, title = {Stochastic simulation efficiency}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus4-31479}, year = {2014}, abstract = {The work described in this report can be broadly divided into two sections. The first section considers two export features. We describe how the export for stochastic Petri nets to SBML level 1 has been added to the Petri net modelling and simulation tool Snoopy. This task was accomplished by making appropriate changes to the existing export code to generate SBML level 2. Also we demonstrate in detail, how the direct export for coloured Petri nets to both levels (i.e. 1 and 2) of SBML was realised. The next section summarises the performed comparison of different stochastic simulation tools for biochemical reaction networks. We first compare BioNetGen and SSC with each other by performing simulations on non-coloured Petri nets. Then, we compare the remaining four tools, i.e. Cain, Marcie, Snoopy and Stochkit with each other by performing simulation on coloured Petri nets. This work builds on results by Aman Sinha [19].}, subject = {Petri Nets; SBML; Stochastic simulation; Biochemical reaction networks; Export; MathML; Stochastisches Modell; Petri-Netz; Export}, language = {en} } @techreport{Nolte2013, author = {Nolte, J{\"o}rg}, title = {12. GI/ITG Fachgespr{\"a}ch Sensornetze}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus-28986}, year = {2013}, abstract = {Drahtlose Sensornetze stellen eine vielversprechende Technologie zur Beobachtung und Beeinflussung von Vorg{\"a}ngen in der realen Welt dar. Autonome Sensorknoten nehmen dabei Parameter der Umwelt durch Sensoren wahr und k{\"o}nnen diese durch Aktoren beeinflussen. Viele solcher autonomen und ressourcenbeschr{\"a}nkten Knoten kooperieren dabei mittels drahtloser Kommunikation. Die Eigenschaften dieser Knoten und Netze implizieren eine Vielzahl von neuartigen Herausforderungen, die sich in einer regen Forschungsaktivit{\"a}t widerspiegeln. Ziel dieser Reihe von Fachgespr{\"a}chen ist es, Wissenschaftlerinnen und Wissenschaftlern aus Hochschule und Industrie die M{\"o}glichkeit zu einem informellen Gedankenaustausch zu geben und die Kooperation in diesem multidisziplin{\"a}ren Forschungsbereich zu verst{\"a}rken.}, subject = {Drahtloses Sensorsystem}, language = {en} } @techreport{NoackSchmittSaretz2013, author = {Noack, Tino and Schmitt, Ingo and Saretz, Sascha}, title = {OVA-based multi-class classification for data stream anomaly detection}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:co1-opus-28187}, year = {2013}, abstract = {Mobile cyber-physical systems (MCPSs), such as the International Space Station, are equipped with sensors which produce sensor data streams. Continuous changes like wear and tear influence the system states of a MCPS continually during runtime. Hence, monitoring is necessary to provide reliability and to avoid critical damage. Although, the monitoring process is limited by resource restrictions. Therefore, the focal point of the present paper is on time-efficient multi-class data stream anomaly detection. Our contribution is bifid. First, we use a one-versusall classification model to combine a set of heterogeneous one-class classifiers consecutively. Such a chain of one-class classifiers provides a very flexible structure while the administrative overhead is reasonably low. Second, based on the classifier chain, we introduce classifier pre-selection.}, subject = {Informatik}, language = {en} }