Florian Sobieczky

Florian Sobieczky
Software Competence Center Hagenberg | SCCH · Data Analysis Systems Group

PhD

About

31
Publications
3,109
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
78
Citations
Introduction
My main interest is the probability theory of random processes associated with graphs and discrete structures (like percolation, random graphs, queues, random walks).
Additional affiliations
August 2016 - present
Software Competence Center Hagenberg
Position
  • Researcher
December 2015 - March 2016
runIT solutions GmbH
Position
  • Data Analyst, Researcher
September 2015 - November 2015
AMS AG
Position
  • Consultant

Publications

Publications (31)
Article
Full-text available
A transformation of unimodal multivariate data is introduced for increased precision in the estimation of the exponential decay type of the underlying density. The transformation renders the contour lines of the probability density function more uniformly spherical and enables conservation of unimodality, without assuming the ability to efficiently...
Chapter
For the problem of parameter estimation in time-series models of a stream of physical measurements under potential local quality impairments (indicated by an additional binary variable), a lamplighter-graph representation is introduced. This representation makes conditional estimation using large samples conditioned on specific degrees of quality i...
Article
The problem of transferring calibrations from a primary to a secondary instrument, that is, calibration transfer (CT), has been a matter of considerable research in chemometrics over the past decades. Current state‐of‐the‐art (SoA) methods like (piecewise) direct standardization perform well when suitable transfer standards are available. However,...
Article
Full-text available
In various industrial production environments a burn-in phase of a specific settling-length precedes a stable (i.e. steady-state, stationary) process mode. The identification of the corresponding physical parameters may be difficult to perform in the presence of strong noise. We propose a method using isotonic regression which circumvents the negat...
Preprint
Full-text available
By means of a local surrogate approach, an analytical method to yield explanations of AI-predictions in the framework of regression models is defined. In the case of the AI-model producing additive corrections to the predictions of a base model, the explanations are delivered in the form of a shift of its interpretable parameters as long as the AI-...
Preprint
Full-text available
This introduces the use of Before and After correction Parameter Comparison (BAPC) as a method in explainable artificial intelligence. It is based on a local surrogate approach with the novel feature of providing interpretations of small AI corrections by parameter shifts of an interpretable base model.
Preprint
Full-text available
In various industrial production environments a burn-in phase of a specific settling-length precedes a stable (i.e. steady-state, stationary) process mode. The identification of the corresponding physical parameters may be difficult to perform in the presence of strong noise. We propose a method using isotonic regression which circumvents the negat...
Article
Full-text available
Using a local surrogate approach from explainable AI, a new prediction method for the performance of start-up companies based on psychological profiles is proposed. The method assumes the existence of an interpreted ‘base model’, the predictions of which are enhanced by an AI-model delivering corrections that improve the overall accuracy. The surro...
Book
This volume constitutes the refereed proceedings of the workshops held at the 32nd International Conference on Database and Expert Systems Applications, DEXA 2021, held in a virtual format in September 2021: The 12th International Workshop on Biological Knowledge Discovery from Data (BIOKDD 2021), the 5th International Workshop on Cyber-Security an...
Article
Full-text available
The main challenges are discussed together with the lessons learned from past and ongoing research along the development cycle of machine learning systems. This will be done by taking into account intrinsic conditions of nowadays deep learning models, data and software quality issues and human-centered artificial intelligence (AI) postulates, inclu...
Chapter
The main challenges along with lessons learned from ongoing research in the application of machine learning systems in practice are discussed, taking into account aspects of theoretical foundations, systems engineering, and human-centered AI postulates. The analysis outlines a fundamental theory-practice gap which superimposes the challenges of AI...
Preprint
Full-text available
The problem of transferring calibrations from a primary to a secondary instrument, i.e. calibration transfer (CT), has been a matter of considerable research in chemometrics over the past decades. Current state-of-the-art (SoA) methods like (piecewise) direct standardization perform well when suitable transfer standards are available. However, stab...
Chapter
For years, the amount of data generated in many industrial production plants has been said to have great potential for improving maintenance processes. In order to leverage this potential in practice, however, it is necessary to overcome a number of hurdles from automated data exchange, linking separate data sources, evaluating the actual data qual...
Article
Full-text available
Inverse problems appear in multiple industrial applications. Solving such inverse problems require the repeated solution of the forward problem. This is the most time-consuming stage when employing inversion techniques, and it constitutes a severe limitation when the inversion needs to be performed in real-time. In here, we focus on the real-time i...
Article
Full-text available
An adaptive linear model predictive control strategy is introduced for the problem of demand side energy management, involving a photovoltaic device, a battery, and a heat pump. Moreover, the heating influence of solar radiation via the glass house effect is considered. Global sunlight radiation intensity and the outside temperature are updated by...
Article
Full-text available
By developing the entropy theory of random walks on equivalence relations and analyzing the asymptotic geometry of horospheric products we describe the Poisson boundary for random walks on random horospheric products of trees.
Article
Using the theory of M/M/1 queues at stationarity, we provide criteria of stability (recurrence) for a stochastic inventory model with an observed selling rate and optimally chosen buying rate. Optimality is based on the maximum gain under stability, where buying and selling-prices, as well as shop- and stock-keeping costs are incorporated into the...
Article
Full-text available
We construct measures invariant with respect to equivalence relations which are graphed by horospheric products of trees. The construction is based on using conformal systems of boundary measures on treed equivalence relations. The existence of such an invariant measure allows us to establish amenability of horospheric products of random trees.
Article
Full-text available
For horocyclic products of percolation subtrees of regular trees, we show almost sure amenability. Under a symmetry condition concerning the growth of the two percolation trees, we show the existence of an increasing Foelner sequence (which we call strong amenability).
Article
Full-text available
Bounds for the expected return probability of the delayed random walk on finite clusters of an invariant percolation on transitive unimodular graphs are derived. They are particularly suited for the case of critical Bernoulli percolation and the associated heavy-tailed cluster size distributions. The upper bound relies on the fact that cartesian pr...
Article
By an eigenvalue comparison-technique(20), the expected return probability of the delayed random walk on critical Bernoulli bond percolation clusters on the two- dimensional Euclidean lattice is estimated. The results are generalised to invariant per- colations on unimodular graphs with almost surely finite clusters. The approach involves using the...
Article
Full-text available
A comparison technique for finite random walks on finite graphs is introduced, using the well-known interlacing method. It yields improved return probability bounds. A key feature is the incorporation of parts of the spectrum of the transition matrix other than just the principal eigenvalue. As an application, an upper bound of the expected return...
Conference Paper
The definition of consonance as the ability to resolve a sound into the pitch categories is introduced. For a vector space of chords a norm is used to evaluate the consonance linearly in dependence of the instrument used. It is shown that in the corresponding Hilbert space the chords which usually appear together in a conventional musical piece are...
Article
By an eigenvalue comparison-technique polynomial bounds for the expected return probability of the delayed random walk on critical Bernoulli bond percolation clusters are derived. The results refer to invariant percolations on unimodular transitive planar graphs with almost surely finite critical clusters. Estimates for the integrated density of st...
Article
The expected n-step return-probability E µ P o X n = o] of a random wal X n with symmetric transition probabilities on a random partial graph of a regular graph G of degree δ with transitive automorphism group Aut(G) is estimated. The law µ of the random edge-set is assumed to be stationary with respect to some transitive, unimodular subgroup Γ of...

Network

Cited By

Projects

Project (1)
Project
The detection, diagnosis, and prediction of anomalies in process data comprises the conventional predictive maintenance program used in production environments. In this project, state of the art machine learning methods are to be equipped with an interpretation in the framework of a physical or probabilistic interpretable 'base-model' allowing for optimised process parameter tuning. The goal is to provide qualitative information about the process from the quantitative corrections supplied by the predictive process analysis.