arXiv subjects

Sebastian Dorn

Publications and source records attributed to Sebastian Dorn.

At least 19 recordsLinked to original sources

arXiv2026

Operationally Guided Placement-Aware Learning for Industrial Online 3D Bin Packing

The online three-dimensional bin packing problem (3D-BPP) is a longstanding challenge in logistics and industrial palletizing. Recent learning-based methods use a learned policy to select among feasible candidate placements. Performance depends on the candidate generator and representation, especially in industrial settings where packings must be space-efficient, stable, compact, and balanced. However, prior work has mainly optimized the policy, while candidate generation and representation remain largely geometry-driven. We address this gap with OPAL, an operationally guided placement-aware learning framework for industrial online 3D-BPP which combines an Operationally Guided Empty-Maximal-Space generator (OG-EMS), an operational representation for each candidate placement, and a masked ranking policy trained with proximal policy optimization. OG-EMS evaluates multiple anchors within each free-space region and prioritizes low, well-supported, compact, and spatially diverse placements. An xLSTM-based Placement Encoder models dependencies among geometric and operational candidate attributes, while a lightweight recurrent core combines the resulting embeddings with the current item and pallet state to rank feasible actions. On the BED-BPP benchmark, OPAL achieves a mean space utilization of 0.49, with improvements of 15.1% from operationally guided candidate generation and 6.3% from learned ranking, while maintaining robust inference-time performance.

cs.AI↗

arXiv2026

GENPACK: KPI-Guided Multi-Criteria Genetic Algorithm for Industrial 3D Bin Packing

The three-dimensional bin packing problem (3D-BPP) is a longstanding challenge in operations research and logistics. While classical heuristics and constructive methods can generate packings efficiently, they often fail to satisfy industrial requirements such as stability, balance, and handling feasibility. Metaheuristics such as genetic algorithms (GAs) offer greater flexibility, but pure GA approaches frequently struggle with efficiency, parameter sensitivity, and scalability to industrial order sizes. These limitations are particularly evident at real-world pallet dimensions, where even state-of-the-art methods often fail to produce robust, deployable solutions. We propose a KPI-guided GA-based pipeline for industrial 3D-BPP that integrates key performance indicators (KPIs) directly into a scalarized fitness function. The method combines a layer-based chromosome representation, domain-specific operators, and constructive heuristics to balance efficiency and feasibility. On the BED-BPP benchmark of 1,500 real-world orders, our GENPACK pipeline consistently outperforms heuristic and learning-based baselines, achieving up to 35% higher space utilization and 15-20% stronger surface support, while exhibiting lower variance across orders. These gains come at a modest runtime cost but remain practical for batch-scale deployment, yielding stable, balanced, and space-efficient packings.

cs.NE↗

arXiv2025

X-TRACK: Physics-Aware xLSTM for Realistic Vehicle Trajectory Prediction

Accurate trajectory prediction is crucial for safe and reliable autonomous driving systems, requiring models that capture long-term temporal dependencies while accounting for social interactions among neighboring vehicles in highway driving scenarios. While Long Short Term Memory (LSTM) networks have been widely used in the domain of trajectory prediction, they have limitations such as limited memory capacity and scalar cell state. The recently introduced Extended Long Short Term Memory (xLSTM) addresses these limitations of traditional LSTMs by introducing exponential gating and enhanced memory structures, making them better suited for modeling long-term temporal dependencies. Despite their potential, xLSTM-based models remain underexplored in the context of vehicle trajectory prediction. This paper introduces a novel xLSTM-based highway trajectory prediction framework, X-TRAJ, as the first application of xLSTM, and its physics-aware variant, X-TRACK (eXtended LSTM for TRAjectory prediction Constraint by Kinematics), which explicitly integrates vehicle motion kinematics into the model learning process. By introducing physical constraints, the proposed model generates realistic and feasible highway trajectories. A comprehensive evaluation on the publicly available highway datasets, highD and NGSIM, demonstrates that X-TRACK outperforms state-of-the-art baselines on highD and is among the state-of-the-art models on the NGSIM dataset.

cs.LG↗

arXiv2024

Reliable Trajectory Prediction and Uncertainty Quantification with Conditioned Diffusion Models

This work introduces the conditioned Vehicle Motion Diffusion (cVMD) model, a novel network architecture for highway trajectory prediction using diffusion models. The proposed model ensures the drivability of the predicted trajectory by integrating non-holonomic motion constraints and physical constraints into the generative prediction module. Central to the architecture of cVMD is its capacity to perform uncertainty quantification, a feature that is crucial in safety-critical applications. By integrating the quantified uncertainty into the prediction process, the cVMD's trajectory prediction performance is improved considerably. The model's performance was evaluated using the publicly available highD dataset. Experiments show that the proposed architecture achieves competitive trajectory prediction accuracy compared to state-of-the-art models, while providing guaranteed drivable trajectories and uncertainty quantification.

cs.LG↗

arXiv2023

Prediction and Interpretation of Vehicle Trajectories in the Graph Spectral Domain

This work provides a comprehensive analysis and interpretation of the graph spectral representation of traffic scenarios. Based on a spatio-temporal vehicle interaction graph, an observed traffic scenario can be transformed into the graph spectral domain by means of the multidimensional Graph Fourier Transformation. Since these spectral scenario representations have shown to successfully incorporate the complex and interactive nature of traffic scenarios, the beneficial feature representation is employed for the purpose of predicting vehicle trajectories. This work introduces GFTNNv2, a deep learning network predicting vehicle trajectories in the graph spectral domain. Evaluation of the GFTNNv2 on the publicly available datasets highD and NGSIM shows a performance gain of up to 25% in comparison to state-of-the-art prediction approaches.

cs.AI↗

arXiv2023

Optimization and Interpretability of Graph Attention Networks for Small Sparse Graph Structures in Automotive Applications

For automotive applications, the Graph Attention Network (GAT) is a prominently used architecture to include relational information of a traffic scenario during feature embedding. As shown in this work, however, one of the most popular GAT realizations, namely GATv2, has potential pitfalls that hinder an optimal parameter learning. Especially for small and sparse graph structures a proper optimization is problematic. To surpass limitations, this work proposes architectural modifications of GATv2. In controlled experiments, it is shown that the proposed model adaptions improve prediction performance in a node-level regression task and make it more robust to parameter initialization. This work aims for a better understanding of the attention mechanism and analyzes its interpretability of identifying causal importance.

cs.LG↗

arXiv2023

Gradient Derivation for Learnable Parameters in Graph Attention Networks

This work provides a comprehensive derivation of the parameter gradients for GATv2 [4], a widely used implementation of Graph Attention Networks (GATs). GATs have proven to be powerful frameworks for processing graph-structured data and, hence, have been used in a range of applications. However, the achieved performance by these attempts has been found to be inconsistent across different datasets and the reasons for this remains an open research question. As the gradient flow provides valuable insights into the training dynamics of statistically learning models, this work obtains the gradients for the trainable model parameters of GATv2. The gradient derivations supplement the efforts of [2], where potential pitfalls of GATv2 are investigated.

cs.LG↗

arXiv2022

SFIP: Coarse-Grained Syscall-Flow-Integrity Protection in Modern Systems

Growing code bases of modern applications have led to a steady increase in the number of vulnerabilities. Control-Flow Integrity (CFI) is one promising mitigation that is more and more widely deployed and prevents numerous exploits. CFI focuses purely on one security domain. That is, transitions between user space and kernel space are not protected by CFI. Furthermore, if user space CFI is bypassed, the system and kernel interfaces remain unprotected, and an attacker can run arbitrary transitions. In this paper, we introduce the concept of syscall-flow-integrity protection (SFIP) that complements the concept of CFI with integrity for user-kernel transitions. Our proof-of-concept implementation relies on static analysis during compilation to automatically extract possible syscall transitions. An application can opt-in to SFIP by providing the extracted information to the kernel for runtime enforcement. The concept is built on three fully-automated pillars: First, a syscall state machine, representing possible transitions according to a syscall digraph model. Second, a syscall-origin mapping, which maps syscalls to the locations at which they can occur. Third, an efficient enforcement of syscall-flow integrity in a modified Linux kernel. In our evaluation, we show that SFIP can be applied to large scale applications with minimal slowdowns. In a micro- and a macrobenchmark, it only introduces an overhead of 13.1% and 1.8%, respectively. In terms of security, we discuss and demonstrate its effectiveness in preventing control-flow-hijacking attacks in real-world applications. Finally, to highlight the reduction in attack surface, we perform an analysis of the state machines and syscall-origin mappings of several real-world applications. On average, SFIP decreases the number of possible transitions by 38.6% compared to seccomp and 90.9% when no protection is applied.

cs.CR↗

arXiv2021

Classification and Uncertainty Quantification of Corrupted Data using Semi-Supervised Autoencoders

Parametric and non-parametric classifiers often have to deal with real-world data, where corruptions like noise, occlusions, and blur are unavoidable - posing significant challenges. We present a probabilistic approach to classify strongly corrupted data and quantify uncertainty, despite the model only having been trained with uncorrupted data. A semi-supervised autoencoder trained on uncorrupted data is the underlying architecture. We use the decoding part as a generative model for realistic data and extend it by convolutions, masking, and additive Gaussian noise to describe imperfections. This constitutes a statistical inference task in terms of the optimal latent space activations of the underlying uncorrupted datum. We solve this problem approximately with Metric Gaussian Variational Inference (MGVI). The supervision of the autoencoder's latent space allows us to classify corrupted data directly under uncertainty with the statistically inferred latent space activations. Furthermore, we demonstrate that the model uncertainty strongly depends on whether the classification is correct or wrong, setting a basis for a statistical "lie detector" of the classification. Independent of that, we show that the generative model can optimally restore the uncorrupted datum by decoding the inferred latent space activations.

cs.LG↗

arXiv2020

Center3D: Center-based Monocular 3D Object Detection with Joint Depth Understanding

Localizing objects in 3D space and understanding their associated 3D properties is challenging given only monocular RGB images. The situation is compounded by the loss of depth information during perspective projection. We present Center3D, a one-stage anchor-free approach, to efficiently estimate 3D location and depth using only monocular RGB images. By exploiting the difference between 2D and 3D centers, we are able to estimate depth consistently. Center3D uses a combination of classification and regression to understand the hidden depth information more robustly than each method alone. Our method employs two joint approaches: (1) LID: a classification-dominated approach with sequential Linear Increasing Discretization. (2) DepJoint: a regression-dominated approach with multiple Eigen's transformations for depth estimation. Evaluating on KITTI dataset for moderate objects, Center3D improved the AP in BEV from $29.7\%$ to $42.8\%$, and the AP in 3D from $18.6\%$ to $39.1\%$. Compared with state-of-the-art detectors, Center3D has achieved the best speed-accuracy trade-off in realtime monocular object detection.

cs.CV↗

arXiv2020

A2D2: Audi Autonomous Driving Dataset

Research in machine learning, mobile robotics, and autonomous driving is accelerated by the availability of high quality annotated data. To this end, we release the Audi Autonomous Driving Dataset (A2D2). Our dataset consists of simultaneously recorded images and 3D point clouds, together with 3D bounding boxes, semantic segmentation, instance segmentation, and data extracted from the automotive bus. Our sensor suite consists of six cameras and five LiDAR units, providing full 360 degree coverage. The recorded data is time synchronized and mutually registered. Annotations are for non-sequential frames: 41,277 frames with semantic segmentation image and point cloud labels, of which 12,497 frames also have 3D bounding box annotations for objects within the field of view of the front camera. In addition, we provide 392,556 sequential frames of unannotated sensor data for recordings in three cities in the south of Germany. These sequences contain several loops. Faces and vehicle number plates are blurred due to GDPR legislation and to preserve anonymity. A2D2 is made available under the CC BY-ND 4.0 license, permitting commercial use subject to the terms of the license. Data and further information are available at https://a2d2-dataset.github.io/.

cs.CV↗

arXiv2019

The primordial magnetic field in our cosmic backyard

We reconstruct the 3D structure of magnetic fields, which were seeded by density perturbations during the radiation dominated epoch of the Universe and later on were evolved by structure formation. To achieve this goal, we rely on three dimensional initial density fields inferred from the 2M++ galaxy compilation via the Bayesian $\texttt{BORG}$ algorithm. Using those, we estimate the magnetogenesis by the so called Harrison mechanism. This effect produced magnetic fields exploiting the different photon drag on electrons and ions in vortical motions, which are exited due to second order perturbation effects in the Early Universe. Subsequently we study the evolution of these seed fields through the non-linear cosmic structure formation by virtue of a MHD simulation to obtain a 3D estimate for the structure of this primordial magnetic field component today. At recombination we obtain a reliable lower limit on the large scale magnetic field strength around $10^{-23} \mathrm{G}$, with a power spectrum peaking at about $ 2\, \mathrm{Mpc}^{-1}h$ in comoving scales. At present we expect this evolved primordial field to have strengthts above $\approx 10^{-27}\, \mathrm{G}$ and $\approx 10^{-29}\, \mathrm{G}$ in clusters of galaxies and voids, respectively. We also calculate the corresponding Faraday rotation measure map and show the magnetic field morphology and strength for specific objects of the Local Universe.

astro-ph.CO↗

arXiv2016

Cosmic expansion history from SNe Ia data via information field theory -- the charm code

We present charm (cosmic history agnostic reconstruction method), a novel inference algorithm that reconstructs the cosmic expansion history as encoded in the Hubble parameter $H(z)$ from SNe Ia data. The novelty of the approach lies in the usage of information field theory, a statistical field theory that is very well suited for the construction of optimal signal recovery algorithms. The charm algorithm infers non-parametrically $s(a)=\ln(ρ(a)/ρ_{\mathrm{crit}0})$, the density evolution which determines $H(z)$, without assuming an analytical form of $ρ(a)$ but only its smoothness with the scale factor $a=(1+z)^{-1}$. The inference problem of recovering the signal $s(a)$ from the data is formulated in a fully Bayesian way. In detail, we have rewritten the signal as the sum of a background cosmology and a perturbation. This allows us to determine the maximum a posteriory estimate of the signal by an iterative Wiener filter method. Applying charm to the Union2.1 supernova compilation, we have recovered a cosmic expansion history that is fully compatible with the standard $Λ$CDM cosmological expansion history with parameter values consistent with the results of the Planck mission.

astro-ph.CO↗

arXiv2015

Stochastic determination of matrix determinants

Matrix determinants play an important role in data analysis, in particular when Gaussian processes are involved. Due to currently exploding data volumes, linear operations - matrices - acting on the data are often not accessible directly but are only represented indirectly in form of a computer routine. Such a routine implements the transformation a data vector undergoes under matrix multiplication. While efficient probing routines to estimate a matrix's diagonal or trace, based solely on such computationally affordable matrix-vector multiplications, are well known and frequently used in signal inference, there is no stochastic estimate for its determinant. We introduce a probing method for the logarithm of a determinant of a linear operator. Our method rests upon a reformulation of the log-determinant by an integral representation and the transformation of the involved terms into stochastic expressions. This stochastic determinant determination enables large-size applications in Bayesian inference, in particular evidence calculations, model comparison, and posterior determination.

physics.data-an↗

arXiv2015

All-sky reconstruction of the primordial scalar potential from WMAP temperature data

An essential quantity required to understand the physics of the early Universe, in particular the inflationary epoch, is the primordial scalar potential $Φ$ and its statistics. We present for the first time an all-sky reconstruction of $Φ$ with corresponding $1σ$-uncertainty from WMAP's cosmic microwave background (CMB) temperature data -- a map of the very early Universe right after the inflationary epoch. This has been achieved by applying a Bayesian inference method that separates the whole inverse problem of the reconstruction into many independent ones, each of them solved by an optimal linear filter (Wiener filter). In this way, the three-dimensional potential $Φ$ gets reconstructed slice by slice resulting in a thick shell of nested spheres around the comoving distance to the last scattering surface. Each slice represents the primordial scalar potential $Φ$ projected onto a sphere with corresponding distance. Furthermore, we present an advanced method for inferring $Φ$ and its power spectrum simultaneously from data, but argue that applying it requires polarization data with high signal-to-noise levels not available yet. Future CMB data should improve results significantly, as polarization data will fill the present $\ell-$blind gaps of the reconstruction.

astro-ph.CO↗

arXiv2015

Signal inference with unknown response: Calibration-uncertainty renormalized estimator

The calibration of a measurement device is crucial for every scientific experiment, where a signal has to be inferred from data. We present CURE, the calibration uncertainty renormalized estimator, to reconstruct a signal and simultaneously the instrument's calibration from the same data without knowing the exact calibration, but its covariance structure. The idea of CURE, developed in the framework of information field theory, is starting with an assumed calibration to successively include more and more portions of calibration uncertainty into the signal inference equations and to absorb the resulting corrections into renormalized signal (and calibration) solutions. Thereby, the signal inference and calibration problem turns into solving a single system of ordinary differential equations and can be identified with common resummation techniques used in field theories. We verify CURE by applying it to a simplistic toy example and compare it against existent self-calibration schemes, Wiener filter solutions, and Markov Chain Monte Carlo sampling. We conclude that the method is able to keep up in accuracy with the best self-calibration methods and serves as a non-iterative alternative to it.

physics.data-an↗

arXiv2014

Generic inference of inflation models by local non-Gaussianity

The presence of multiple fields during inflation might seed a detectable amount of non-Gaussianity in the curvature perturbations, which in turn becomes observable in present data sets like the cosmic microwave background (CMB) or the large scale structure (LSS). Within this proceeding we present a fully analytic method to infer inflationary parameters from observations by exploiting higher-order statistics of the curvature perturbations. To keep this analyticity, and thereby to dispense with numerically expensive sampling techniques, a saddle-point approximation is introduced whose precision has been validated for a numerical toy example. Applied to real data, this approach might enable to discriminate among the still viable models of inflation.

astro-ph.CO↗

arXiv2014

Generic inference of inflation models by non-Gaussianity and primordial power spectrum reconstruction

We present a generic inference method for inflation models from observational data by the usage of higher-order statistics of the curvature perturbation on uniform density hypersurfaces. This method is based on the calculation of the posterior for the primordial non-Gaussianity parameters $f_\text{NL}$ and $g_\text{NL}$, which in general depend on specific parameters of inflation and reheating models, and enables to discriminate among the still viable inflation models. To keep analyticity as far as possible to dispense with numerically expensive sampling techniques a saddle-point approximation is introduced, whose precision is validated for a numerical toy example. The mathematical formulation is done in a generic way so that the approach remains applicable to cosmic microwave background data as well as to large scale structure data. Additionally, we review a few currently interesting inflation models and present numerical toy examples thereof in two and three dimensions to demonstrate the efficiency of the higher-order statistics method. A second quantity of interest is the primordial power spectrum. Here, we present two Bayesian methods to infer it from observational data, the so called critical filter and an extension thereof with smoothness prior, both allowing for a non-parametric spectrum reconstruction. These methods are able to reconstruct the spectra of the observed perturbations and the primordial ones of curvature perturbation even in case of non-Gaussianity and partial sky coverage. We argue that observables like $T-$ and $B-$modes permit to measure both spectra. This also allows to infer the level of non-Gaussianity generated since inflation.

astro-ph.CO↗