arXiv subjects

Sebastian Mueller

Publications and source records attributed to Sebastian Mueller.

15 recordsLinked to original sources

arXiv2025

Beyond Pretty Pictures: Combined Single- and Multi-Image Super-resolution for Sentinel-2 Images

Super-resolution aims to increase the resolution of satellite images by reconstructing high-frequency details, which go beyond na\"ive upsampling. This has particular relevance for Earth observation missions like Sentinel-2, which offer frequent, regular coverage at no cost; but at coarse resolution. Its pixel footprint is too large to capture small features like houses, streets, or hedge rows. To address this, we present SEN4X, a hybrid super-resolution architecture that combines the advantages of single-image and multi-image techniques. It combines temporal oversampling from repeated Sentinel-2 acquisitions with a learned prior from high-resolution Pl\'eiades Neo data. In doing so, SEN4X upgrades Sentinel-2 imagery to 2.5 m ground sampling distance. We test the super-resolved images on urban land-cover classification in Hanoi, Vietnam. We find that they lead to a significant performance improvement over state-of-the-art super-resolution baselines.

eess.IV↗

arXiv2025

Masked Conditioning for Deep Generative Models

Datasets in engineering domains are often small, sparsely labeled, and contain numerical as well as categorical conditions. Additionally. computational resources are typically limited in practical applications which hinders the adoption of generative models for engineering tasks. We introduce a novel masked-conditioning approach, that enables generative models to work with sparse, mixed-type data. We mask conditions during training to simulate sparse conditions at inference time. For this purpose, we explore the use of various sparsity schedules that show different strengths and weaknesses. In addition, we introduce a flexible embedding that deals with categorical as well as numerical conditions. We integrate our method into an efficient variational autoencoder as well as a latent diffusion model and demonstrate the applicability of our approach on two engineering-related datasets of 2D point clouds and images. Finally, we show that small models trained on limited data can be coupled with large pretrained foundation models to improve generation quality while retaining the controllability induced by our conditioning scheme.

cs.LG↗

arXiv2025

MeshFleet: Filtered and Annotated 3D Vehicle Dataset for Domain Specific Generative Modeling

Generative models have recently made remarkable progress in the field of 3D objects. However, their practical application in fields like engineering remains limited since they fail to deliver the accuracy, quality, and controllability needed for domain-specific tasks. Fine-tuning large generative models is a promising perspective for making these models available in these fields. Creating high-quality, domain-specific 3D datasets is crucial for fine-tuning large generative models, yet the data filtering and annotation process remains a significant bottleneck. We present MeshFleet, a filtered and annotated 3D vehicle dataset extracted from Objaverse-XL, the most extensive publicly available collection of 3D objects. Our approach proposes a pipeline for automated data filtering based on a quality classifier. This classifier is trained on a manually labeled subset of Objaverse, incorporating DINOv2 and SigLIP embeddings, refined through caption-based analysis and uncertainty estimation. We demonstrate the efficacy of our filtering method through a comparative analysis against caption and image aesthetic score-based techniques and fine-tuning experiments with SV3D, highlighting the importance of targeted data selection for domain-specific 3D generative modeling.

cs.CV↗

arXiv2024

GeoBiked: A Dataset with Geometric Features and Automated Labeling Techniques to Enable Deep Generative Models in Engineering Design

We provide a dataset for enabling Deep Generative Models (DGMs) in engineering design and propose methods to automate data labeling by utilizing large-scale foundation models. GeoBiked is curated to contain 4 355 bicycle images, annotated with structural and technical features and is used to investigate two automated labeling techniques: The utilization of consolidated latent features (Hyperfeatures) from image-generation models to detect geometric correspondences (e.g. the position of the wheel center) in structural images and the generation of diverse text descriptions for structural images. GPT-4o, a vision-language-model (VLM), is instructed to analyze images and produce diverse descriptions aligned with the system-prompt. By representing technical images as Diffusion-Hyperfeatures, drawing geometric correspondences between them is possible. The detection accuracy of geometric points in unseen samples is improved by presenting multiple annotated source images. GPT-4o has sufficient capabilities to generate accurate descriptions of technical images. Grounding the generation only on images leads to diverse descriptions but causes hallucinations, while grounding it on categorical labels restricts the diversity. Using both as input balances creativity and accuracy. Successfully using Hyperfeatures for geometric correspondence suggests that this approach can be used for general point-detection and annotation tasks in technical images. Labeling such images with text descriptions using VLMs is possible, but dependent on the models detection capabilities, careful prompt-engineering and the selection of input information. Applying foundation models in engineering design is largely unexplored. We aim to bridge this gap with a dataset to explore training, finetuning and conditioning DGMs in this field and suggesting approaches to bootstrap foundation models to process technical images.

cs.CV↗

arXiv2021

The relentless variability of Mrk 421 from the TeV to the radio

The origin of the gamma-ray emission of the blazar Mrk 421 is still a matter of debate. We used 5.5 years of unbiased observing campaign data, obtained using the FACT telescope and the Fermi LAT detector at TeV and GeV energies, the longest and densest so far, together with contemporaneous multi-wavelength observations, to characterise the variability of Mrk 421 and to constrain the underlying physical mechanisms. We studied and correlated light curves obtained by ten different instruments and found two significant results. The TeV and X-ray light curves are very well correlated with a lag of <0.6 days. The GeV and radio (15 Ghz band) light curves are widely and strongly correlated. Variations of the GeV light curve lead those in the radio. Lepto-hadronic and purely hadronic models in the frame of shock acceleration predict proton acceleration or cooling timescales that are ruled out by the short variability timescales and delays observed in Mrk 421. Instead the observations match the predictions of leptonic models.

astro-ph.HE↗

arXiv2019

Robustness and efficiency of leaderless probabilistic consensus protocols within Byzantine infrastructures

This paper investigates leaderless binary majority consensus protocols with low computational complexity in noisy Byzantine infrastructures. Using computer simulations, we show that explicit randomization of the consensus protocol can significantly increase the robustness towards faulty and malicious nodes. We identify the optimal amount of randomness for various Byzantine attack strategies on different kinds of network topologies.

cs.DC↗

arXiv2019

Quality analysis in acyclic production networks

The production network under examination consists of a number of workstations. Each workstation is a parallel configuration of machines performing the same kind of tasks on a given part. Parts move from one workstation to another and at each workstation a part is assigned randomly to a machine. We assume that the production network is acyclic, that is, a part does not return to a workstation where it previously received service. Furthermore, we assume that the quality of the end product is additive, that is, the sum of the quality contributions of the machines along the production path. The contribution of each machine is modeled by a separate random variable. Our main result is the construction of estimators that allow pairwise and multiple comparison of the means and variances of machines in the same workstation. These comparisons then may lead to the identification of unreliable machines. We also discuss the asymptotic distributions of the estimators that allow the use of standard statistical tests and decision making.

stat.AP↗

arXiv2017

Regularity of the speed of biased random walk in a one-dimensional percolation model

We consider biased random walks on the infinite cluster of a conditional bond percolation model on the infinite ladder graph. Axelsson-Fisk and H\"aggstr\"om established for this model a phase transition for the asymptotic linear speed $\overline{\mathrm{v}}$ of the walk. Namely, there exists some critical value $\lambda_{\mathrm{c}}>0$ such that $\overline{\mathrm{v}}>0$ if $\lambda\in (0,\lambda_{\mathrm{c}})$ and $\overline{\mathrm{v}}=0$ if $\lambda>\lambda_{\mathrm{c}}$. We show that the speed $\overline{\mathrm{v}}$ is continuous in $\lambda$ on the interval $(0,\lambda_{\mathrm{c}})$ and differentiable on $(0,\lambda_{\mathrm{c}}/2)$. Moreover, we characterize the derivative as a covariance. For the proof of the differentiability of $\overline{\mathrm{v}}$ on $(0,\lambda_{\mathrm{c}}/2)$, we require and prove a central limit theorem for the biased random walk. Additionally, we prove that the central limit theorem fails to hold for $\lambda \geq \lambda_{\mathrm{c}}/2$.

math.PR↗

arXiv2015

Infinite excursions of rotor walks on regular trees

A rotor configuration on a graph contains in every vertex an infinite ordered sequence of rotors, each is pointing to a neighbor of the vertex. After sampling a configuration according to some probability measure, a rotor walk is a deterministic process: at each step it chooses the next unused rotor in its current location, and uses it to jump to the neighboring vertex to which it points. Rotor walks capture many aspects of the expected behavior of simple random walks. However, this similarity breaks down for the property of having an infinite excursion. In this paper we study that question for natural random configuration models on regular trees. Our results suggest that in this context the rotor model behaves like the simple random walk unless it is not "close to" the standard rotor-router model.

math.PR↗

arXiv2014

Rotor-routing on Galton-Watson trees

A rotor-router walk on a graph is a deterministic process, in which each vertex is endowed with a rotor that points to one of the neighbors. A particle located at some vertex first rotates the rotor in a prescribed order, and then it is routed to the neighbor the rotor is now pointing at. In the current work we make a step toward in understanding the behavior of rotor-router walks on random trees. More precisely, we consider random i.i.d. initial configurations of rotors on Galton-Watson trees, i.e. on a family tree arising from a Galton-Watson process, and give a classification in recurrence and transience for rotor-router walks on these trees.

math.PR↗

arXiv2013

Renewal theory for random walks on surface groups

We construct a renewal structure for random walks on surface groups. The renewal times are defined as times when the random walks enters a particular type of a cone and never leaves it again. As a consequence, the trajectory of the random walk can be expressed as an "aligned union" of i.i.d. trajectories between the renewal times. Once having established this renewal structure, we prove a central limit theorem for the distance to the origin under exponential moment conditions. Analyticity of the speed and of the asymptotic variance are natural consequences of our approach. Furthermore, our method applies to groups with infinitely many ends and therefore generalizes classic results on central limit theorems on free groups.

math.PR↗

arXiv2011

The 5 hr pulse period and broadband spectrum of the Symbiotic X-ray Binary 3A 1954+319

We present an analysis of the highly variable accreting X-ray pulsar 3A 1954+319 using 2005-2009 monitoring data obtained with INTEGRAL and Swift. This considerably extends the pulse period history and covers flaring episodes in 2005 and 2008. In 2006 the source was identified as one of only a few known symbiotic X-ray binaries (SyXBs), i.e., systems composed of a neutron star accreting from the inhomogeneous medium around an M-giant star. The extremely long pulse period of 5.3 hr is directly visible in the 2008 INTEGRAL-ISGRI outburst light curve. The pulse profile is double peaked and generally not significantly energy dependent although there is an indication of possible softening during the main pulse. During the outburst a strong spin-up of -1.8 10^(-4) hr hr^(-1) occurred. Between 2005 and 2008 a long-term spin-down trend of 2.1 10^-5 hr hr^(-1) was observed for the first time for this source. The 3-80 keV pulse peak spectrum of 3A 1954+319 during the 2008 flare could be well described by a thermal Comptonization model. We interpret the results within the framework of a recently developed quasi-spherical accretion model for SyXBs.

astro-ph.HE↗

arXiv2011

Sublimation Temperature of Circumstellar Dust Particles and Its Importance for Dust Ring Formation

Dust particles in orbit around a star drift toward the central star by the Poynting-Robertson effect and pile up by sublimation. We analytically derive the pile-up magnitude, adopting a simple model for optical cross sections. As a result, we find that the sublimation temperature of drifting dust particles plays the most important role in the pile-up rather than their optical property does. Dust particles with high sublimation temperature form a significant dust ring, which could be found in the vicinity of the sun through in-situ spacecraft measurements. While the existence of such a ring in a debris disk could not be identified in the spectral energy distribution (SED), the size of a dust-free zone shapes the SED. Since we analytically obtain the location and temperature of sublimation, these analytical formulae are useful to find such sublimation evidences.

astro-ph.EP↗

arXiv2010

The cold origin of the warm dust around epsilon Eridani

Context: The K2V star eps Eri hosts one known inner planet, an outer Kuiper belt analog, and an inner disk of warm dust. Spitzer/IRS measurements indicate that the warm dust is present at distances as close as a few AU from the star. Its origin is puzzling, since an "asteroid belt" that could produce this dust would be unstable because of the known inner planet. Aims: Here we test the hypothesis that the observed warm dust is generated by collisions in the outer belt and is transported inward by Poynting-Robertson (P-R) drag and strong stellar winds. Methods: We simulated a steady-state distribution of dust particles outside 10AU with a collisional code and in the inner region (r<10AU) with single-particle numerical integrations. By assuming homogeneous spherical dust grains composed of water ice and silicate, we calculated the thermal emission of the dust and compared it with observations. We investigated two different orbital configurations for the inner planet inferred from RV measurements, one with a highly eccentric orbit of e=0.7 and another one with a moderate one of e=0.25. We also produced a simulation without a planet. Results: Our models can reproduce the shape and magnitude of the observed SED from mid-IR to sub-mm wavelengths, as well as the Spitzer/MIPS radial brightness profiles. The best-fit dust composition includes both ice and silicates. The results are similar for the two possible planetary orbits and without a planet. Conclusions: The observed warm dust in the system can indeed stem from the outer belt and be transported inward by P-R and stellar wind drag. The inner planet has little effect on the distribution of dust, so that the planetary orbit could not be constrained. Reasonable agreement between the model and observations can only be achieved by relaxing the assumption of purely silicate dust and assuming a mixture of silicate and ice in comparable amounts.

astro-ph.EP↗

arXiv2005

The critical Branching Markov Chain is transient

We investigate recurrence and transience of Branching Markov Chains (BMC) in discrete time. Branching Markov Chains are clouds of particles which move (according to an irreducible underlying Markov Chain) and produce offspring independently. The offspring distribution can depend on the location of the particle. If the offspring distribution is constant for all locations, these are Tree-Indexed Markov chains in the sense of \cite{benjamini94}. Starting with one particle at location $x$, we denote by $\alpha(x)$ the probability that $x$ is visited infinitely often by the cloud. Due to the irreducibility of the underlying Markov Chain, there are three regimes: either $\alpha(x) = 0$ for all $x$ (transient regime), or $0 < \alpha(x) < 1$ for all $x$ (weakly recurrent regime) or $\alpha(x) = 1$ for all $x$ (strongly recurrent regime). We give classification results, including a sufficient condition for transience in the general case. If the mean of the offspring distribution is constant, we give a criterion for transience involving the spectral radius of the underlying Markov Chain and the mean of the offspring distribution.

math.PR↗