Inst. Softwaretechnik und Theoretische Informatik

277 Items

Recent Submissions
Efficient Bayesian Inference of Sigmoidal Gaussian Cox Processes

Donner, Christian ; Opper, Manfred (2018)

We present an approximate Bayesian inference approach for estimating the intensity of a inhomogeneous Poisson process, where the intensity function is modelled using a Gaussian process (GP) prior via a sigmoid link function. Augmenting the model using a latent marked Poisson process and Polya--Gamma random variables we obtain a representation of the likelihood which is conjugate to the GP prior...

Advanced EEG signal processing with applications in brain-computer interfaces

Nicolae, Irina-Emilia (2019)

Advances in signal processing push forward the Neurotechnology domain along with the Brain-Computer Interface (BCI) research which deals with the analysis of brain activity. Heading for a future that will most probably happen, where either healthy persons or people with disabilities communicate and control external devices without muscle control, a symbiotic relationship between humans and mach...

Bayesian inference of inhomogeneous point process models

Donner, Christian (2019)

Arrival times of airplanes, positions of car accidents or astronomical objects in space, locations of ecological crisis, spike times of neurons, etc. are all data that surround us and can be viewed as realisations of point processes. Nowadays, the modelling of these data becomes increasingly more important, when we attempt to draw meaningful conclusions from this ever expanding amount of data. ...

Compiling Modelica

Höger, Christoph (2019)

The equation-based object-oriented modeling language Modelica is an openly accessible standard with many implementations and applications. Most, if not all, tools that execute a Modelica simulation follow a common scheme: Models are loaded, composed, analyzed, and transformed into a system of equations. This system is then further simplified, translated into efficient code, and simulated....

Optimizing the depth and the direction of prospective planning using information values

Sezener, Can Eren ; Dezfouli, Amir ; Keramati, Mehdi (2019-03-12)

Evaluating the future consequences of actions is achievable by simulating a mental search tree into the future. Expanding deep trees, however, is computationally taxing. Therefore, machines and humans use a plan-until-habit scheme that simulates the environment up to a limited depth and then exploits habitual values as proxies for consequences that may arise in the future. Two outstanding quest...

Features and machine learning systems for structured and sequential data

Schwenk, Guido (2019)

Modern web and communication technology relies heavily on sequential and structured data for its process execution and communication protocols. Due to its complex properties, a manual analysis and detection of problems on this data is too time-consuming and expensive, and hence not feasible. As a consequence, features and automatic learning systems on this type of data are highly sought after. ...

Exploring Deployment Strategies for the Tor Network

Döpmann, Christoph ; Rust, Sebastian ; Tschorsch, Florian (2018)

In response to upcoming performance and security challenges of anonymity networks like Tor, it will be of crucial importance to be able to develop and deploy performance improvements and state-of-the-art countermeasures. In this paper, we therefore explore different deployment strategies and review their applicability, impact, and risks to the Tor network. In a simulation-based evaluation, whic...

Towards a Concurrent and Distributed Route Selection for Payment Channel Networks

Rohrer, Elias ; Laß, Jann-Frederik ; Tschorsch, Florian (2017)

Payment channel networks use off-chain transactions to provide virtually arbitrary transaction rates. In this paper, we provide a new perspective on payment channels and consider them as a flow network. We propose an extended push-relabel algorithm to find payment flows in a payment channel network. Our algorithm enables a distributed and concurrent execution without violating capacity constrai...

Webchain: Verifiable Citations and References for the World Wide Web

Rohrer, Elias ; Heidel, Steffen ; Tschorsch, Florian (2018)

Readers’ capability to consider and assess sources is imperative. Digital preservation efforts, however, mostly neglected citation provenance, which is a necessity for transparent source verification. We therefore present Webchain, a new system enabling verifiable citations and references on the World Wide Web. Its architecture combines a distributed ledger with secure timestamping to ensure hi...

CircuitStart: A Slow Start For Multi-Hop Anonymity Systems

Döpmann, Christoph ; Tschorsch, Florian (2018)

In order to improve the performance of anonymity networks like Tor, custom transport protocols have been proposed to efficiently deal with the multi-hop nature of such overlay networks. In this work, we tackle the issue of quickly, but safely, ramping up the congestion window during the initial phase of a circuit's lifetime. We propose a tailored startup mechanism called CircuitStart that trans...

P2KMV: A Privacy-preserving Counting Sketch for Efficient and Accurate Set Intersection Cardinality Estimations

Sparka, Hagen ; Tschorsch, Florian ; Scheuermann, Björn (2018-05-01)

In this paper, we propose P2KMV, a novel privacy-preserving counting sketch, based on the k minimum values algorithm. With P2KMV, we offer a versatile privacy-enhanced technology for obtaining statistics, following the principle of data minimization, and aiming for the sweet spot between privacy, accuracy, and computational efficiency. As our main contribution, we develop methods to perform set...

Exploring Deployment Strategies for the Tor Network [Extended Version]

Döpmann, Christoph ; Rust, Sebastian ; Tschorsch, Florian (2018-07-07)

In response to upcoming performance and security challenges of anonymity networks like Tor, it will be of crucial importance to be able to develop and deploy performance improvements and state-of-the-art countermeasures. In this paper, we therefore explore different deployment strategies and review their applicability to the Tor network. In particular, we consider flag day, dual stack, translat...

Representations and optimizations for embedded parallel dataflow languages

Alexandrov, Alexander (2019)

Parallel dataflow engines such as Apache Hadoop, Apache Spark, and Apache Flink have emerged as an alternative to relational databases more suitable for the needs of modern data analysis applications. One of the main characteristics of these systems is their scalable programming model, based on distributed collections and parallel transformations. Notable examples are Flink’s DataSet and Spark’...

One-class classification in the presence of point, collective, and contextual anomalies

Görnitz, Nico (2019)

Anomaly detection has a prominent position in the processing pipeline of any real-world data-driven application. Its central goal is to detect and separate valid data points from malicious-anomalous-ones such that the cleaned data set can be processed further. In many applications, anomalies are even the prime objects of interest and need to be exposed early in order to avoid loss, e.g. in cred...

Brain-Computer Interface - Motor Imagery Data

Blankertz, Benjamin ; Vidaurre, Carmen ; Sannelli, Claudia ; Kübler, Andrea ; Halder, Sebastian ; Hammer, Eva-Maria (2019-01)

We provide a data set of a BCI study using a motor imagery paradigm. In a calibration session, participants were instructed by cues to perform different types of imagined movements. The pair of classes resulting in the most promising discrimination was chosen and a classifier was trained. That classifier was used in the feedback session to let the participants move a cursor horizontally accordi...

Opening the machine learning black box with Layer-wise Relevance Propagation

Lapuschkin, Sebastian (2019)

Machine learning techniques such as (Deep) Neural Networks are successfully solving a plethora of tasks, e.g. in image recognition and text analysis, and provide novel predictive models for complex physical, biological and chemical systems. However, due to the nested complex and non-linear structure of many machine learning models, this comes with the disadvantage of them acting as a black box,...

Computational modeling of glutamate-induced calcium signal generation and propagation in astrocytes

Oschmann, Franziska (2018)

Since the 1990s researchers have shown that astrocytes generate calcium oscillations in response to neuronal activity and propagate them as intercellular calcium waves over long distances. Moreover, astrocytes release transmitters in a calcium-dependent manner and by that signal to neurons. These discoveries have made astrocytes and especially calcium signal generation and propagation in astroc...

A statistical physics approach to inference problems on random networks

Bachschmid Romano, Ludovica (2018)

Recent advances in measurement technologies have resulted in the availability of large datasets from a variety of fields spanning the natural and social sciences. This posed the challenge to develop new statistical tools to extract relevant information from the data. A paradigmatic model that has been successfully applied to analyze large datasets is the Ising model of binary spins interacting ...

Investigating the effects of weak extracellular fields on single neurons: a modelling approach

Aspart, Florian (2018)

In the past decades, the rise of transcranial current stimulation (tCS) has sparkled an increasing interest in the effects of weak extracellular electric fields on neural activity. These fields, such as induced during tCS, have been shown to polarize the neuronal membrane and, consequently, to modulate the spiking activity. In this thesis, I follow a modelling approach to investigate how single...

Low dimensional visualization and modelling of data using distance-based models

Grünhage, Gina (2018)

This thesis consists of two parts, which seek low-dimensional representations for visualization and analysis of data. Both parts use rather different types of models and inference methods. In both cases, however, the models show inherent invariances, which need to be coped with during the optimization procedures. The first part addresses a fundamental problem in machine learning, namely, the ch...