Know Center Research

Publikationen

Geiger Bernhard, Schuppler Barbara
2023
Given the development of automatic speech recognition based techniques for creating phonetic annotations of large speech corpora, there has been a growing interest in investigating the frequencies of occurrence of phonological and reduction processes. Given that most studies have analyzed these processes separately, they did not provide insights about their cooccurrences. This paper contributes with introducing graph theory methods for the analysis of pronunciation variation in a large corpus of Austrian German conversational speech. More specifically, we investigate how reduction processes that are typical for spontaneous German in general co-occur with phonological processes typical for the Austrian German variety. Whereas our concrete findings are of special interest to scientists investigating variation in German, the approach presented opens new possibilities to analyze pronunciation variation in large corpora of different speaking styles in any language.
Geiger Bernhard, Jahani Alireza, Hussain Hussain, Groen Derek
2023
In this work, we investigate Markov aggregation for agent-based models (ABMs). Specifically, if the ABM models agent movements on a graph, if its ruleset satisfies certain assumptions, and if the aim is to simulate aggregate statistics such as vertex populations, then the ABM can be replaced by a Markov chain on a comparably small state space. This equivalence between a function of the ABM and a smaller Markov chain allows to reduce the computational complexity of the agent-based simulation from being linear in the number of agents, to being constant in the number of agents and polynomial in the number of locations.We instantiate our theory for a recent ABM for forced migration (Flee). We show that, even though the rulesets of Flee violate some of our necessary assumptions, the aggregated Markov chain-based model, MarkovFlee, achieves comparable accuracy at substantially reduced computational cost. Thus, MarkovFlee can help NGOs and policy makers forecast forced migration in certain conflict scenarios in a cost-effective manner, contributing to fast and efficient delivery of humanitarian relief.
Posch Stefan, Gößnitzer Clemens, Rohrhofer Franz Martin, Geiger Bernhard, Wimmer Andreas
2023
The turbulent jet ignition concept using prechambers is a promising solution to achieve stable combustion at lean conditions in large gas engines, leading to high efficiency at low emission levels. Due to the wide range of design and operating parameters for large gas engine prechambers, the preferred method for evaluating different designs is computational fluid dynamics (CFD), as testing in test bed measurement campaigns is time-consuming and expensive. However, the significant computational time required for detailed CFD simulations due to the complexity of solving the underlying physics also limits its applicability. In optimization settings similar to the present case, i.e., where the evaluation of the objective function(s) is computationally costly, Bayesian optimization has largely replaced classical design-of-experiment. Thus, the present study deals with the computationally efficient Bayesian optimization of large gas engine prechambers design using CFD simulation. Reynolds-averaged-Navier-Stokes simulations are used to determine the target values as a function of the selected prechamber design parameters. The results indicate that the chosen strategy is effective to find a prechamber design that achieves the desired target values.
Rohrhofer Franz Martin, Posch Stefan, Gößnitzer Clemens, García-Oliver José M., Geiger Bernhard
2023
Flamelet models are widely used in computational fluid dynamics to simulate thermochemical processes in turbulent combustion. These models typically employ memory-expensive lookup tables that are predetermined and represent the combustion process to be simulated.Artificial neural networks (ANNs) offer a deep learning approach that can store this tabular data using a small number of network weights, potentially reducing the memory demands of complex simulations by orders of magnitude.However, ANNs with standard training losses often struggle with underrepresented targets in multivariate regression tasks, e.g., when learning minor species mass fractions as part of lookup tables.This paper seeks to improve the accuracy of an ANN when learning multiple species mass fractions of a hydrogen (\ce{H2}) combustion lookup table. We assess a simple, yet effective loss weight adjustment that outperforms the standard mean-squared error optimization and enables accurate learning of all species mass fractions, even of minor species where the standard optimization completely fails. Furthermore, we find that the loss weight adjustment leads to more balanced gradients in the network training, which explains its effectiveness.
Adilova Linara, Geiger Bernhard, Fischer Asja
2023
The information-theoretic framework promises to explain the predictive power of neural networks. In particular, the information plane analysis, which measures mutual information (MI) between input and representation as well as representation and output, should give rich insights into the training process. This approach, however, was shown to strongly depend on the choice of estimator of the MI. The problem is amplified for deterministic networks if the MI between input and representation is infinite. Thus, the estimated values are defined by the different approaches for estimation, but do not adequately represent the training process from an information-theoretic perspective. In this work, we show that dropout with continuously distributed noise ensures that MI is finite. We demonstrate in a range of experiments that this enables a meaningful information plane analysis for a class of dropout neural networks that is widely used in practice.
Berger Katharina, Rusch Magdalena, Pohlmann Antonia, Popowicz Martin, Geiger Bernhard, Gursch Heimo, Schöggl Josef-Peter, Baumgartner Rupert J.
2023
Digital product passports (DPPs) are an emerging technology and are considered as enablers of sustainable and circular value chains as they support sustainable product management (SPM) by gathering and containing product life cycle data. However, some life cycle data are considered sensitive by stakeholders, resulting in a reluctance to share such data. This contribution provides a concept illustrating how data science and machine learning approaches enable electric vehicle battery (EVB) value chain stakeholders to carry out confidentiality-preserving data exchange via a DPP. This, in turn, can support overcoming data sharing reluctances, consequently facilitating sustainability data management on a DPP for an EVB. The concept development comprised a literature review to identify data needs for sustainable EVB management, data management challenges, and potential data science approaches for data management support. Furthermore, three explorative focus group workshops and follow-up consultations with data scientists were conducted to discuss identified data sciences approaches. This work complements the emerging literature on digitalization and SPM by exploring the specific potential of data science, and machine learning approaches enabling sustainability data management and reducing data sharing reluctance. Furthermore, practical relevance is given, as this concept may provide practitioners with new impulses regarding DPP development and implementation.
Xue Yani, Li Miqing, Arabnejad Hamid, Suleimenova, Geiger Bernhard, Jahani Alireza, Groen Derek
2022
In the context of humanitarian support for forcibly displaced persons, camps play an important role in protecting people and ensuring their survival and health. A challenge in this regard is to find optimal locations for establishing a new asylum-seeker/unrecognized refugee or IDPs (internally displaced persons) camp. In this paper we formulate this problem as an instantiation of the well-known facility location problem (FLP) with three objectives to be optimized. In particular, we show that AI techniques and migration simulations can be used to provide decision support on camp placement.
De Freitas Joao Pedro, Berg Sebastian, Geiger Bernhard, Mücke Manfred
2022
In this paper, we frame homogeneous-feature multi-task learning (MTL) as a hierarchical representation learning problem, with one task-agnostic and multiple task-specific latent representations. Drawing inspiration from the information bottleneck principle and assuming an additive independent noise model between the task-agnostic and task-specific latent representations, we limit the information contained in each task-specific representation. It is shown that our resulting representations yield competitive performance for several MTL benchmarks. Furthermore, for certain setups, we show that the trained parameters of the additive noise model are closely related to the similarity of different tasks. This indicates that our approach yields a task-agnostic representation that is disentangled in the sense that its individual dimensions may be interpretable from a task-specific perspective.
Steger Sophie, Geiger Bernhard, Smieja Marek
2022
We connect the problem of semi-supervised clustering to constrained Markov aggregation, i.e., the task of partitioning the state space of a Markov chain. We achieve this connection by considering every data point in the dataset as an element of the Markov chain's state space, by defining the transition probabilities between states via similarities between corresponding data points, and by incorporating semi-supervision information as hard constraints in a Hartigan-style algorithm. The introduced Constrained Markov Clustering (CoMaC) is an extension of a recent information-theoretic framework for (unsupervised) Markov aggregation to the semi-supervised case. Instantiating CoMaC for certain parameter settings further generalizes two previous information-theoretic objectives for unsupervised clustering. Our results indicate that CoMaC is competitive with the state-of-the-art
Schweimer Christoph, Gfrerer Christine, Lugstein Florian, Pape David, Velimsky Jan, Elsässer Robert, Geiger Bernhard
2022
Online social networks are a dominant medium in everyday life to stay in contact with friends and to share information. In Twitter, users can connect with other users by following them, who in turn can follow back. In recent years, researchers studied several properties of social networks and designed random graph models to describe them. Many of these approaches either focus on the generation of undirected graphs or on the creation of directed graphs without modeling the dependencies between reciprocal (i.e., two directed edges of opposite direction between two nodes) and directed edges. We propose an approach to generate directed social network graphs that creates reciprocal and directed edges and considers the correlation between the respective degree sequences.Our model relies on crawled directed graphs in Twitter, on which information w.r.t.\ a topic is exchanged or disseminated. While these graphs exhibit a high clustering coefficient and small average distances between random node pairs (which is typical in real-world networks), their degree sequences seem to follow a $\chi^2$-distribution rather than power law. To achieve high clustering coefficients, we apply an edge rewiring procedure that preserves the node degrees.We compare the crawled and the created graphs, and simulate certain algorithms for information dissemination and epidemic spreading on them. The results show that the created graphs exhibit very similar topological and algorithmic properties as the real-world graphs, providing evidence that they can be used as surrogates in social network analysis. Furthermore, our model is highly scalable, which enables us to create graphs of arbitrary size with almost the same properties as the corresponding real-world networks.
Geiger Bernhard
2021
(extended abstract)
Geiger Bernhard, Al-Bashabsheh Ali
2021
We derive two sufficient conditions for a function of a Markov random field (MRF) on a given graph to be a MRF on the same graph. The first condition is information-theoretic and parallels a recent information-theoretic characterization of lumpability of Markov chains. The second condition, which is easier to check, is based on the potential functions of the corresponding Gibbs field. We illustrate our sufficient conditions at the hand of several examples and discuss implications for practical applications of MRFs. As a side result, we give a partial characterization of functions of MRFs that are information preserving.
Klimashevskaia Anastasia, Geiger Bernhard, Hagmüller Martin, Helic Denis, Fischer Frank
2020
(extended abstract)
Hobisch Elisbeth, Scholger Martina, Fuchs Alexandra, Geiger Bernhard, Koncar Philipp, Saric Sanja
2020
(extended abstract)
Schrunner Stefan, Geiger Bernhard, Zernig Anja, Kern Roman
2020
Classification has been tackled by a large number of algorithms, predominantly following a supervised learning setting. Surprisingly little research has been devoted to the problem setting where a dataset is only partially labeled, including even instances of entirely unlabeled classes. Algorithmic solutions that are suited for such problems are especially important in practical scenarios, where the labelling of data is prohibitively expensive, or the understanding of the data is lacking, including cases, where only a subset of the classes is known. We present a generative method to address the problem of semi-supervised classification with unknown classes, whereby we follow a Bayesian perspective. In detail, we apply a two-step procedure based on Bayesian classifiers and exploit information from both a small set of labeled data in combination with a larger set of unlabeled training data, allowing that the labeled dataset does not contain samples from all present classes. This represents a common practical application setup, where the labeled training set is not exhaustive. We show in a series of experiments that our approach outperforms state-of-the-art methods tackling similar semi-supervised learning problems. Since our approach yields a generative model, which aids the understanding of the data, it is particularly suited for practical applications.
Gogolenko Sergiy, Groen Derek, Suleimenova Dian, Mahmood Imra, Lawenda Marcin, Nieto De Santos Javie, Hanley Joh, Vukovic Milana, Kröll Mark, Geiger Bernhard, Elsaesser Rober, Hoppe Dennis
2020
Accurate digital twinning of the global challenges (GC) leadsto computationally expensive coupled simulations. These simulationsbring together not only different models, but also various sources of mas-sive static and streaming data sets. In this paper, we explore ways tobridge the gap between traditional high performance computing (HPC)and data-centric computation in order to provide efficient technologicalsolutions for accurate policy-making in the domain of GC. GC simula-tions in HPC environments give rise to a number of technical challengesrelated to coupling. Being intended to reflect current and upcoming situ-ation for policy-making, GC simulations extensively use recent streamingdata coming from external data sources, which requires changing tradi-tional HPC systems operation. Another common challenge stems fromthe necessity to couple simulations and exchange data across data centersin GC scenarios. By introducing a generalized GC simulation workflow,this paper shows commonality of the technical challenges for various GCand reflects on the approaches to tackle these technical challenges in theHiDALGO project
Fuchs Alexandra, Geiger Bernhard, Hobisch Elisabeth, Koncar Philipp, More Jacqueline, Saric Sanja, Scholger Martina
2020
Fuchs Alexandra, Geiger Bernhard, Hobisch Elisabeth, Koncar Philipp, Saric Sanja, Scholger Martina
2019
with contributions from Denis Helic and Jacqueline More
Lindstaedt Stefanie , Geiger Bernhard, Pirker Gerhard
2019
Big Data and data-driven modeling are receiving more and more attention in various research disciplines, where they are often considered as universal remedies. Despite their remarkable records of success, in certain cases a purely data-driven approach has proven to be suboptimal or even insufficient.This extended abstract briefly defines the terms Big Data and data-driven modeling and characterizes scenarios in which a strong focus on data has proven to be promising. Furthermore, it explains what progress can be made by fusing concepts from data science and machine learning with current physics-based concepts to form hybrid models, and how these can be applied successfully in the field of engine pre-simulation and engine control.
Schweimer Christoph, Geiger Bernhard, Suleimenova Diana, Groen Derek, Gfrerer Christine, Pape David, Elsaesser Robert, Kocsis Albert Tihamér, Liszkai B., Horváth Zoltan
2019
Toller Maximilian, Geiger Bernhard, Kern Roman
2019
Distance-based classification is among the most competitive classification methods for time series data. The most critical componentof distance-based classification is the selected distance function.Past research has proposed various different distance metrics ormeasures dedicated to particular aspects of real-world time seriesdata, yet there is an important aspect that has not been considered so far: Robustness against arbitrary data contamination. In thiswork, we propose a novel distance metric that is robust against arbitrarily “bad” contamination and has a worst-case computationalcomplexity of O(n logn). We formally argue why our proposedmetric is robust, and demonstrate in an empirical evaluation thatthe metric yields competitive classification accuracy when appliedin k-Nearest Neighbor time series classification.
Geiger Bernhard
2019
joint work with Tobias Koch, Universidad Carlos III de Madrid
Maritsch Martin, Diana Suleimenova, Geiger Bernhard, Derek Groen
2019
Geiger Bernhard, Schrunner Stefan, Kern Roman
2019
Schrunner and Geiger have contributed equally to this work.