TY  - THES
A1  - Kaminski, Roland
T1  - Complex reasoning with answer set programming
N2  - Answer Set Programming (ASP) allows us to address knowledge-intensive search and optimization problems in a declarative way due to its integrated modeling, grounding, and solving workflow. A problem is modeled using a rule based language and then grounded and solved. Solving results in a set of stable models that correspond to solutions of the modeled problem. In this thesis, we present the design and implementation of the clingo system---perhaps, the most
widely used ASP system. It features a rich modeling language originating from the field of knowledge representation and reasoning, efficient grounding algorithms based on database evaluation techniques, and high performance solving algorithms based on Boolean satisfiability (SAT) solving technology.
The contributions of this thesis lie in the design of the modeling language, the design and implementation of the grounding algorithms, and the design and implementation of an Application Programmable Interface (API) facilitating the use of ASP in real world applications and the implementation of complex forms of reasoning beyond the traditional ASP workflow.
KW  - Answer Set Programming
KW  - Declarative Problem Solving
KW  - Grounding Theory
KW  - Preference Handling
KW  - Answer Set Solving modulo Theories
KW  - Temporal Answer Set Solving
Y1  - 2023
ER  - 
TY  - GEN
A1  - Prasse, Paul
A1  - Iversen, Pascal
A1  - Lienhard, Matthias
A1  - Thedinga, Kristina
A1  - Herwig, Ralf
A1  - Scheffer, Tobias
T1  - Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity Prediction
T2  - Zweitveröffentlichungen der Universität Potsdam : Mathematisch-Naturwissenschaftliche Reihe
N2  - Large-scale databases that report the inhibitory capacities of many combinations of candidate drug compounds and cultivated cancer cell lines have driven the development of preclinical drug-sensitivity models based on machine learning. However, cultivated cell lines have devolved from human cancer cells over years or even decades under selective pressure in culture conditions. Moreover, models that have been trained on in vitro data cannot account for interactions with other types of cells. Drug-response data that are based on patient-derived cell cultures, xenografts, and organoids, on the other hand, are not available in the quantities that are needed to train high-capacity machine-learning models. We found that pre-training deep neural network models of drug sensitivity on in vitro drug-sensitivity databases before fine-tuning the model parameters on patient-derived data improves the models’ accuracy and improves the biological plausibility of the features, compared to training only on patient-derived data. From our experiments, we can conclude that pre-trained models outperform models that have been trained on the target domains in the vast majority of cases.
T3  - Zweitveröffentlichungen der Universität Potsdam : Mathematisch-Naturwissenschaftliche Reihe - 1300 
KW  - deep neural networks
KW  - drug-sensitivity prediction
KW  - anti-cancer drugs
Y1  - 2023
U6  - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:517-opus4-577341
SN  - 1866-8372
SP  - 1
EP  - 14
PB  - Universitätsverlag Potsdam
CY  - Potsdam
ER  - 
TY  - JOUR
A1  - Prasse, Paul
A1  - Iversen, Pascal
A1  - Lienhard, Matthias
A1  - Thedinga, Kristina
A1  - Herwig, Ralf
A1  - Scheffer, Tobias
T1  - Pre-Training on In Vitro and Fine-Tuning on Patient-Derived Data Improves Deep Neural Networks for Anti-Cancer Drug-Sensitivity Prediction
JF  - MDPI
N2  - Large-scale databases that report the inhibitory capacities of many combinations of candidate drug compounds and cultivated cancer cell lines have driven the development of preclinical drug-sensitivity models based on machine learning. However, cultivated cell lines have devolved from human cancer cells over years or even decades under selective pressure in culture conditions. Moreover, models that have been trained on in vitro data cannot account for interactions with other types of cells. Drug-response data that are based on patient-derived cell cultures, xenografts, and organoids, on the other hand, are not available in the quantities that are needed to train high-capacity machine-learning models. We found that pre-training deep neural network models of drug sensitivity on in vitro drug-sensitivity databases before fine-tuning the model parameters on patient-derived data improves the models’ accuracy and improves the biological plausibility of the features, compared to training only on patient-derived data. From our experiments, we can conclude that pre-trained models outperform models that have been trained on the target domains in the vast majority of cases.
KW  - deep neural networks
KW  - drug-sensitivity prediction
KW  - anti-cancer drugs
Y1  - 2022
U6  - https://doi.org/10.3390/cancers14163950
SN  - 2072-6694
VL  - 14
SP  - 1
EP  - 14
PB  - MDPI
CY  - Basel, Schweiz
ET  - 16
ER  - 
TY  - JOUR
A1  - Stede, Manfred
T1  - From connectives to coherence relations
BT  - a case study of German contrastrive connectives
JF  - Revue roumaine de linguistique : RRL = Romanian review of linguistics
N2  - The notion of coherence relations is quite widely accepted in general, but concrete proposals differ considerably on the questions of how they should be motivated, which relations are to be assumed, and how they should be defined. This paper takes a "bottom-up" perspective by assessing the contribution made by linguistic signals (connectives), using insights from the relevant literature as well as verification by practical text annotation. We work primarily with the German language here and focus on the realm of contrast. Thus, we suggest a new inventory of contrastive connective functions and discuss their relationship to contrastive coherence relations that have been proposed in earlier work.
KW  - coherence relation
KW  - connective
KW  - contrast
KW  - concession
KW  - corpus analysis
Y1  - 2020
SN  - 0035-3957
VL  - 65
IS  - 3
SP  - 213
EP  - 233
PB  - Ed. Academiei Române
CY  - Bucureşti
ER  - 
TY  - GEN
A1  - Mühlbauer, Felix
A1  - Schröder, Lukas
A1  - Skoncej, Patryk
A1  - Schölzel, Mario
T1  - Handling manufacturing and aging faults with software-based techniques in tiny embedded systems
T2  - 18th IEEE Latin American Test Symposium (LATS 2017)
N2  - Non-volatile memory area occupies a large portion of the area of a chip in an embedded system. Such memories are prone to manufacturing faults, retention faults, and aging faults. The paper presents a single software based technique that allows for handling all of these fault types in tiny embedded systems without the need for hardware support. This is beneficial for low-cost embedded systems with simple memory architectures. A software infrastructure and a flow are presented that demonstrate how the presented technique is used in general for fault handling right after manufacturing and in-the-field. Moreover, a full implementation is presented for a MSP430 microcontroller, along with a discussion of the performance, overhead, and reliability impacts.
Y1  - 2027
SN  - 978-1-5386-0415-1
U6  - https://doi.org/10.1109/LATW.2017.7906756
PB  - IEEE
CY  - New York
ER  - 
TY  - GEN
A1  - Mühlbauer, Felix
A1  - Schröder, Lukas
A1  - Schölzel, Mario
T1  - On hardware-based fault-handling in dynamically scheduled processors
T2  - 20th International Symposium on Design and Diagnostics of Electronic Circuits & Systems (DDECS) 2017
N2  - This paper describes architectural extensions for a dynamically scheduled processor, so that it can be used in three different operation modes, ranging from high-performance, to high-reliability. With minor hardware-extensions of the control path, the resources of the superscalar data-path can be used either for high-performance execution, fail-safe-operation, or fault-tolerant-operation. This makes the processor-architecture a very good candidate for applications with dynamically changing reliability requirements, e.g. for automotive applications. The paper reports the hardware-overhead for the extensions, and investigates the performance penalties introduced by the fail-safe and fault-tolerant mode. Furthermore, a comprehensive fault simulation was carried out in order to investigate the fault-coverage of the proposed approach.
Y1  - 2017
SN  - 978-1-5386-0472-4
U6  - https://doi.org/10.1109/DDECS.2017.7934572
SN  - 2334-3133
SN  - 2473-2117
SP  - 201
EP  - 206
PB  - IEEE
CY  - New York
ER  - 
TY  - JOUR
A1  - Everardo Pérez, Flavio Omar
A1  - Osorio, Mauricio
T1  - Towards an answer set programming methodology for constructing programs following a semi-automatic approach
BT  - extended and revised version
JF  - Electronic notes in theoretical computer science
N2  - Answer Set Programming (ASP) is a successful rule-based formalism for modeling and solving knowledge-intense combinatorial (optimization) problems. Despite its success in both academic and industry, open challenges like automatic source code optimization, and software engineering remains. This is because a problem encoded into an ASP might not have the desired solving performance compared to an equivalent representation. Motivated by these two challenges, this paper has three main contributions. First, we propose a developing process towards a methodology to implement ASP programs, being faithful to existing methods. Second, we present ASP encodings that serve as the basis from the developing process. Third, we demonstrate the use of ASP to reverse the standard solving process. That is, knowing answer sets in advance, and desired strong equivalent properties, “we” exhaustively reconstruct ASP programs if they exist. This paper was originally motivated by the search of propositional formulas (if they exist) that represent the semantics of a new aggregate operator. Particularly, a parity aggregate. This aggregate comes as an improvement from the already existing parity (xor) constraints from xorro, where lacks expressiveness, even though these constraints fit perfectly for reasoning modes like sampling or model counting. To this end, this extended version covers the fundaments from parity constraints as well as the xorro system. Hence, we delve a little more in the examples and the proposed methodology over parity constraints. Finally, we discuss our results by showing the only representation available, that satisfies different properties from the classical logic xor operator, which is also consistent with the semantics of parity constraints from xorro.
KW  - answer set programming
KW  - combinatorial optimization problems
KW  - parity aggregate operator
Y1  - 2020
U6  - https://doi.org/10.1016/j.entcs.2020.10.004
SN  - 1571-0661
VL  - 354
SP  - 29
EP  - 44
PB  - Elsevier
CY  - Amsterdam [u.a.]
ER  - 
TY  - JOUR
A1  - Lorenz, Claas
A1  - Clemens, Vera Elisabeth
A1  - Schrötter, Max
A1  - Schnor, Bettina
T1  - Continuous verification of network security compliance
JF  - IEEE transactions on network and service management
N2  - Continuous verification of network security compliance is an accepted need. Especially, the analysis of stateful packet filters plays a central role for network security in practice. But the few existing tools which support the analysis of stateful packet filters are based on general applicable formal methods like Satifiability Modulo Theories (SMT) or theorem prover and show runtimes in the order of minutes to hours making them unsuitable for continuous compliance verification. In this work, we address these challenges and present the concept of state shell interweaving to transform a stateful firewall rule set into a stateless rule set. This allows us to reuse any fast domain specific engine from the field of data plane verification tools leveraging smart, very fast, and domain specialized data structures and algorithms including Header Space Analysis (HSA). First, we introduce the formal language FPL that enables a high-level human-understandable specification of the desired state of network security. Second, we demonstrate the instantiation of a compliance process using a verification framework that analyzes the configuration of complex networks and devices - including stateful firewalls - for compliance with FPL policies. Our evaluation results show the scalability of the presented approach for the well known Internet2 and Stanford benchmarks as well as for large firewall rule sets where it outscales state-of-the-art tools by a factor of over 41.
KW  - Security
KW  - Tools
KW  - Network security
KW  - Engines
KW  - Benchmark testing;
KW  - Analytical models
KW  - Scalability
KW  - Network
KW  - security
KW  - compliance
KW  - formal
KW  - verification
Y1  - 2021
U6  - https://doi.org/10.1109/TNSM.2021.3130290
SN  - 1932-4537
VL  - 19
IS  - 2
SP  - 1729
EP  - 1745
PB  - Institute of Electrical and Electronics Engineers
CY  - New York
ER  - 
TY  - GEN
A1  - Fabian, Benjamin
A1  - Baumann, Annika
A1  - Ehlert, Mathias
A1  - Ververis, Vasilis
A1  - Ermakova, Tatiana
T1  - CORIA - Analyzing internet connectivity risks using network graphs
T2  - 2017 IEEE International Conference on Communications (ICC)
N2  - The Internet can be considered as the most important infrastructure for modern society and businesses. A loss of Internet connectivity has strong negative financial impacts for businesses and economies. Therefore, assessing Internet connectivity, in particular beyond their own premises and area of direct control, is of growing importance in the face of potential failures, accidents, and malicious attacks. This paper presents CORIA, a software framework for an easy analysis of connectivity risks based on large network graphs. It provides researchers, risk analysts, network managers and security consultants with a tool to assess an organization's connectivity and paths options through the Internet backbone, including a user-friendly and insightful visual representation of results. CORIA is flexibly extensible in terms of novel data sets, graph metrics, and risk scores that enable further use cases. The performance of CORIA is evaluated by several experiments on the Internet graph and further randomly generated networks.
KW  - risk analysis
KW  - connectivity
KW  - graph analysis
KW  - complex networks
KW  - Internet
Y1  - 2017
SN  - 978-1-4673-8999-0
SN  - 978-1-4673-9000-2
U6  - https://doi.org/10.1109/ICC.2017.7996828
SN  - 1550-3607
PB  - IEEE
CY  - Piscataway
ER  - 
TY  - JOUR
A1  - Hempel, Sabrina
A1  - Adolphs, Julian
A1  - Landwehr, Niels
A1  - Willink, Dilya
A1  - Janke, David
A1  - Amon, Thomas
T1  - Supervised machine learning to assess methane emissions of a dairy building with natural ventilation
JF  - Applied Sciences
N2  - A reliable quantification of greenhouse gas emissions is a basis for the development of adequate mitigation measures. Protocols for emission measurements and data analysis approaches to extrapolate to accurate annual emission values are a substantial prerequisite in this context. We systematically analyzed the benefit of supervised machine learning methods to project methane emissions from a naturally ventilated cattle building with a concrete solid floor and manure scraper located in Northern Germany. We took into account approximately 40 weeks of hourly emission measurements and compared model predictions using eight regression approaches, 27 different sampling scenarios and four measures of model accuracy. Data normalization was applied based on median and quartile range. A correlation analysis was performed to evaluate the influence of individual features. This indicated only a very weak linear relation between the methane emission and features that are typically used to predict methane emission values of naturally ventilated barns. It further highlighted the added value of including day-time and squared ambient temperature as features. The error of the predicted emission values was in general below 10%. The results from Gaussian processes, ordinary multilinear regression and neural networks were least robust. More robust results were obtained with multilinear regression with regularization, support vector machines and particularly the ensemble methods gradient boosting and random forest. The latter had the added value to be rather insensitive against the normalization procedure. In the case of multilinear regression, also the removal of not significantly linearly related variables (i.e., keeping only the day-time component) led to robust modeling results. We concluded that measurement protocols with 7 days and six measurement periods can be considered sufficient to model methane emissions from the dairy barn with solid floor with manure scraper, particularly when periods are distributed over the year with a preference for transition periods. Features should be normalized according to median and quartile range and must be carefully selected depending on the modeling approach.
KW  - greenhouse gas
KW  - on-farm evaluation
KW  - emission factor
KW  - regression
KW  - ensemble methods
KW  - gradient boosting
KW  - random forest
KW  - neural networks
KW  - support vector machines
Y1  - 2020
U6  - https://doi.org/10.3390/app10196938
SN  - 2076-3417
VL  - 10
IS  - 19
PB  - MDPI
CY  - Basel
ER  -