Source author record

Jürgen Pilz

Jürgen Pilz appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Machine Learning Applications Methodology

Catalog footprint

What is connected

3works

3topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2021arXiv

A User-Guided Bayesian Framework for Ensemble Feature Selection in Life Science Applications (UBayFS)

Feature selection represents a measure to reduce the complexity of high-dimensional datasets and gain insights into the systematic variation in the data. This aspect is of specific importance in domains that rely on model interpretability, such as life sciences. We propose UBayFS, an ensemble feature selection technique embedded in a Bayesian statistical framework. Our approach considers two sources of information: data and domain knowledge. We build a meta-model from an ensemble of elementary feature selectors and aggregate this information in a multinomial likelihood. The user guides UBayFS by weighting features and penalizing specific feature blocks or combinations, implemented via a Dirichlet-type prior distribution and a regularization term. In a quantitative evaluation, we demonstrate that our framework (a) allows for a balanced trade-off between user knowledge and data observations, and (b) achieves competitive performance with state-of-the-art methods.

preprint2021arXiv

From a Point Cloud to a Simulation Model: Bayesian Segmentation and Entropy based Uncertainty Estimation for 3D Modelling

The 3D modelling of indoor environments and the generation of process simulations play an important role in factory and assembly planning. In brownfield planning cases existing data are often outdated and incomplete especially for older plants, which were mostly planned in 2D. Thus, current environment models cannot be generated directly on the basis of existing data and a holistic approach on how to build such a factory model in a highly automated fashion is mostly non-existent. Major steps in generating an environment model in a production plant include data collection and pre-processing, object identification as well as pose estimation. In this work, we elaborate a methodical workflow, which starts with the digitalization of large-scale indoor environments and ends with the generation of a static environment or simulation model. The object identification step is realized using a Bayesian neural network capable of point cloud segmentation. We elaborate how the information on network uncertainty generated by a Bayesian segmentation framework can be used in order to build up a more accurate environment model. The steps of data collection and point cloud segmentation as well as the resulting model accuracy are evaluated on a real-world data set collected at the assembly line of a large-scale automotive production plant. The segmentation network is further evaluated on the publicly available Stanford Large-Scale 3D Indoor Spaces data set. The Bayesian segmentation network clearly surpasses the performance of the frequentist baseline and allows us to increase the accuracy of the model placement in a simulation scene considerably.

preprint2014arXiv

A statistical noise model for a class of Physically Unclonable Functions

The interest in "Physically Unclonable Function"-devices has increased rapidly over the last few years, as they have several interesting properties for system security related applications like, for example, the management of cryptographic keys. Unfortunately, the output provided by these devices is noisy and needs to be corrected for these applications. Related error correcting mechanisms are typically constructed on the basis of an equal error probability for each output bit. This assumption does not hold for Physically Unclonable Functions, where varying error probabilities can be observed. This results in a generalized binomial distribution for the number of errors in the output. The intention of this paper is to discuss a novel Bayesian statistical model for the noise of an especially wide-spread class of Physically Unclonable Functions, which properly handles the varying output stability and also reflects the different noise behaviors observed in a collection of such devices. Furthermore, we compare several different methods for estimating the model parameters and apply the proposed model to concrete measurements obtained within the CODES research project in order to evaluate typical correction and stabilization approaches.