arXiv subjects

Stefan Huber

Publications and source records attributed to Stefan Huber.

At least 19 recordsLinked to original sources

arXiv2026

Exploring Zero-Shot Foundation Models for Multivariate Time Series Anomaly Detection

Multivariate Time Series Anomaly Detection (MTSAD) is essential for reliability and safety in domains such as industrial process monitoring and financial risk management, yet conventional approaches rely on application-specific models that are costly to train and hard to scale. Foundation Models (FMs), pre-trained on broad data with strong zero-shot generalization, have recently become available for univariate time series forecasting, raising the question of whether they can address MTSAD without task-specific training. We investigate the zero-shot application of a univariate forecasting FM, TimesFM, to industrial MTSAD on the Secure Water Treatment (SWaT) benchmark, evaluating two strategies: treating the FM as a per-feature forecaster with thresholded prediction errors, and as an embedder whose intermediate representations feed standard outlier detectors. Neither of our proposed setups is competitive with established baselines; embeddings reveal only partial separation between normal and anomalous segments, insufficient for reliable detection. The cause is that the FM is too effective at capturing temporal dynamics, yielding low error even within fully anomalous windows, so persistent anomalies become indistinguishable from normal behavior. However, these observations yield valuable insights: the error peaks at anomaly boundaries, indicating FMs reliably detect distribution changes. We conclude that the proposed naive zero-shot FMs are unsuitable for MTSAD but promising for change-point detection.

cs.LG↗

arXiv2026

Safe Reinforcement Learning using Ideas from Model Predictive Control

Reinforcement learning (RL) enables the synthesis of control policies directly from data, making it highly appealing for complex cyber-physical systems (CPSs) and robotics. A persistent challenge, however, is ensuring strict, hard safety constraints during the active learning phase. In real-world physical systems, violating mechanical limits can cause irreversible damage, necessitating that exploration remains strictly within safe operational regions. We propose a generalized framework that combines the adaptive, high-performance nature of deep reinforcement learning (DRL) with the formal safety guarantees of model predictive control (MPC). Using a mathematical model of the system dynamics, offline MPC computations define a feasible state-action space, representing all safe combinations of system states and control inputs that guarantee constraint satisfaction. During training and deployment, the RL agent's instantaneous actions are projected onto this globally verified feasible set via a safety filter. We systematically evaluate our generalized approach on a non-linear 1-DoF laboratory testbed, demonstrating successful exploration and stable policy convergence on physical hardware.

cs.LG↗

arXiv2026

Integrating Physics-Informed Neural Networks for Safe Reinforcement Learning in a 1-DoF Helicopter System

Deep reinforcement learning (DRL) offers powerful control for industrial cyber-physical systems (ICPSs), but its "black-box" exploration risks violating strict hardware safety limits. Typically, these constraints are managed through complex reward shaping. In this work-in-progress paper, we embed a differentiable physics model directly into the proximal policy optimization (PPO) actor loss function. By simulating short-horizon future trajectories during training, the policy is penalized for anticipated safety violations independent of the task-reward signal. Evaluated on a simulated 1-degree-of-freedom helicopter testbed with strict pitch constraints, our physics-informed soft regularizations substantially reduce constraint violations while maintaining reliable target tracking.

cs.LG↗

arXiv2026

Anticipatory Reinforcement Learning for Trajectory Tracking

Deep reinforcement learning (DRL) in industrial control often suffers from lag and overshoot due to purely reactive control based on the current tracking error. To achieve anticipatory control without high computational overhead, we introduce a predictive formulation that augments the DRL state space with target velocities and future reference horizons. Evaluating eight configurations using proximal policy optimization (PPO) on a 1-degree-of-freedom (1-DoF) helicopter testbed, simulation results showed a 9-fold error reduction, lowering the mean absolute deviation from 2.73{\deg} to 0.31{\deg}. However, zero-shot transfer to physical hardware revealed a sim-to-real gap. Interestingly, a simpler configuration using a single, further look-ahead horizon matched the real-world top performance of the most complex model (1.11{\deg}). Overall, evaluating various combinations of prediction horizons and target velocities demonstrated that highly granular predictive data is not necessarily required for physical transfer.

cs.LG↗

arXiv2026

Sample-Efficient Pareto Front Modeling for Energy-Aware Reinforcement Learning Using Bayesian Optimization

Industrial automation increasingly demands control strategies that balance operational performance with strict energy efficiency requirements. A common approach to solving this multi-objective problem, particularly within the framework of reinforcement learning (RL), is to formulate a single, scalar reward function that linearly combines the competing objectives. However, the manual weighting of these different objectives is heavily reliant on domain intuition, incredibly time-consuming, prone to human bias, and frequently fails to uncover optimal trade-off solutions. This work addresses the critical challenge of automating the weight selection process to systematically and efficiently discover the Pareto front of optimal trade-off policies. We formulate the weight selection process as a multi-objective Bayesian optimization (MOBO) problem and evaluate its sample efficiency against a standard uniform grid search baseline. Using a physical Quanser Aero 2 testbed configured for 1-DoF pitch control, our results demonstrate that the MOBO approach, utilizing the expected hypervolume improvement (qEHVI) acquisition function, consistently outperforms uniform grid sampling. MOBO achieves superior hypervolume and maximum spread, successfully identifying high-quality, diverse trade-off policies with a reduced evaluation budget, thereby enabling highly efficient energy-aware control in complex mechatronic systems.

cs.LG↗

arXiv2026

Anomaly Detection for Sparse and Irregular Multivariate Time Series with Latent SDEs

Multivariate time series anomaly detection (MTSAD) is critical for a wide range of application areas, such as industrial monitoring, cybersecurity, or healthcare. Real-world data is often sparse, irregularly sampled or partially observed, yet existing methods assume uniformly sampled time series. We propose a generative approach based on Latent SDEs that projects the observed time series on a continuous-time stochastic dynamical system, directly being able to handle missing observations and irregular sampling, while also naturally capturing possible cyclic behavior that many real-world use cases inherently possess. Experiments on six anomaly benchmark datasets show that our proposed method ranks first among state-of-the-art baselines. We further demonstrate that our method remains robust under severe data sparsity, while performance significantly degrades for the tested baseline methods. These results highlight latent SDEs as a natural inductive bias for anomaly detection in multivariate time series, especially in presence of real-world irregularities.

cs.LG↗

arXiv2026

Reconstructing OPC UA Address Spaces from Time-Series Databases

OPC UA has become the dominant open protocol in operational technology. Time-series databases routinely archive OPC UA telemetry but discard the semantic metadata (node hierarchy, engineering units, and type definitions) which gives sensor values their meaning. Recovering this information from a time-series database is non-trivial: namespace indices recorded at the source are session-local and unstable across restarts, and naive merging across multiple source servers results in identifier collisions. We present opcua-ts, an implemented architecture that persists this semantic information alongside its telemetry in a general-purpose time-series database under a lifecycle-stable join key, and that reconstructs the source address space as a live OPC UA endpoint. We characterize the conditions under which the reconstruction is sound across multi-source deployments and validate the approach with a NodeSet2 XML round-trip against the source server. Initial results from a boiler-simulator round-trip indicate that the approach is feasible.

cs.DB↗

arXiv2026

Federated Learning for Multivariate Time Series Anomaly Detection in Industrial Automation

Federated learning (FL) has broadened the horizon for multivariate time series anomaly detection (MTSAD). However, benchmarking such anomaly detection methods within FL paradigm poses data-centric challenges. The existing datasets do not counteract these challenges since they do not simultaneously provide sufficient scale, accurate labels, and freedom from common flaws. In addition, the role of cyclic process behavior, which is common in discrete industrial automation, remains underexplored for MTSAD for the current state of research. This paper aims to shed more light on the literature and address these gaps by introducing a dataset designed with cyclic dynamics arising from the repetitive nature of discrete automation processes and evaluates selected MTSAD methods on both the proposed dataset and a public benchmark dataset.

cs.LG↗

arXiv2026

Chebyshev Policies and the Mountain Car Problem: Reinforcement Learning for Low-Dimensional Control Tasks

We analytically solve the Mountain Car problem, a canonical benchmark in RL, and derive an optimal control solution, closing a gap after 36 years. This enables us to reveal two surprising insights: The optimal control is quite simple, yet modern RL agents display a large gap to optimality. Motivated by the analysis of the optimal control, we introduce Chebyshev policies as a universal (i.e. dense) class of RL policies from first principles. They can be trained as drop-in replacements of neural nets, reducing the regret by a factor of 6.18, while requiring 277 times fewer parameters, fostering sample efficiency, explainability and realtime capability. Chebyshev policies are evaluated on further RL tasks, including a real-world nonlinear motion control testbed. They consistently improve performance over neural nets with PPO, ARS and REINFORCE. Our results demonstrate how Chebyshev policies offer a compelling and lightweight alternative or addition to neural nets for low-dimensional control tasks.

cs.LG↗

arXiv2026

Reinforcement Learning for Optimal Experiment Design in Parameter Identification of Mechatronic Systems

Informative excitation signals are critical for accurate system identification of mechatronic systems, yet classical system identification (SI) approaches require expert knowledge and hand-crafted signal design to respect hardware safety constraints, limiting their generalizability. We propose a reinforcement learning (RL) agent that learns optimal excitation signals for a Quanser Aero 2 testbed while autonomously enforcing safety constraints through reward shaping. Evaluated across 10 independent training seeds, our comprehensive agent achieves competitive estimation accuracy across all three identified parameters, outperforming classical baselines while incurring only 0.75% safety violations.

cs.RO↗

arXiv2026

Generative Modeling of Approximately Periodic Time Series by a Posterior-Weighted Gaussian Process

Discrete automated processes in industrial and cyber-physical systems often exhibit a repetitive structure in which successive repetitions follow a common trajectory while differing in duration, amplitude, and fine-scale dynamics. Such \emph{approximately periodic} behavior poses a challenge for Gaussian Processes (GP) modeling: strictly periodic models suppress inter-repetition variability, while non-periodic models fail to capture the strong structural regularities required for generation. In this work, we propose a stochastic generative model for approximately periodic time series. The model is based on a GP whose posterior is modulated by a novel kernel. Our approach decouples intra-repetition structure from inter-repetition variability through a two-stage construction which yields a generative distribution with a identical mean function across repetitions, while allowing smooth variation between repetitions. The modeling choices are supported by an implementation in which realistic synthetic trajectories are generated from toy datasets.

stat.ML↗

arXiv2026

Finding Patient Zero via Low-Dimensional Geometric Embeddings

We study the patient zero problem in epidemic spreading processes in the independent cascade model and propose a geometric approach for source reconstruction. Using Johnson-Lindenstrauss projections, we embed the contact network into a low-dimensional Euclidean space and estimate the infection source as the node closest to the center of gravity of infected nodes. Simulations on Erd\H{o}s-R\'enyi graphs demonstrate that our estimator achieves meaningful reconstruction accuracy despite operating on compressed observations.

cs.CG↗

arXiv2026

Topologically Stable Hough Transform

We propose an alternative formulation of the well-known Hough transform to detect lines in point clouds. Replacing the discretized voting scheme of the classical Hough transform by a continuous score function, its persistent features in the sense of persistent homology give a set of candidate lines. We also devise and implement an algorithm to efficiently compute these candidate lines.

cs.CG↗

arXiv2026

Two Stages of Folding: Convergent Mechanisms in AI Protein Folding Trunks

How do protein structure prediction models fold proteins? We investigate this question through causal interventions on the folding trunks of ESMFold, OpenFold, and Boltz-1. Across all three models, we find a shared two-stage computational structure. In the first stage, early blocks initialize pairwise biochemical signals: features like charge propagate from sequence into pairwise representations through architecture-specific pathways. In the second stage, late blocks develop pairwise spatial features: distance and contact information accumulate in the pairwise representation. We verify these mechanisms causally by showing that steering charge and distance features induces predictable structural changes. Furthermore, these representations are functionally interchangeable: pairwise states can be linearly aligned and substituted across models. Together, these results suggest that folding trunks with different architectures, inputs, and training procedures converge on a shared representational organization for mapping sequence chemistry into spatial geometry.

cs.LG↗

arXiv2026

Secure Data Bridging in Industry 4.0: An OPC UA Aggregation Approach for Including Insecure Legacy Systems

The increased connectivity of industrial networks has led to a surge in cyberattacks, emphasizing the need for cybersecurity measures tailored to the specific requirements of industrial systems. Modern Industry 4.0 technologies, such as OPC UA, offer enhanced resilience against these threats. However, widespread adoption remains limited due to long installation times, proprietary technology, restricted flexibility, and formal process requirements (e.g. safety certifications). Consequently, many systems do not yet implement these technologies, or only partially. This leads to the challenge of dealing with so-called brownfield systems, which are often placed in isolated security zones to mitigate risks. However, the need for data exchange between secure and insecure zones persists. This paper reviews existing solutions to address this challenge by analysing their approaches, advantages, and limitations. Building on these insights, we identify three key concepts, evaluate their suitability and compatibility, and ultimately introduce the SigmaServer, a novel TCP-level aggregation method. The developed proof-of-principle implementation is evaluated in an operational technology (OT) testbed, demonstrating its applicability and effectiveness in bridging secure and insecure zones.

cs.CR↗

arXiv2025

Embedding the MLOps Lifecycle into OT Reference Models

Machine Learning Operations (MLOps) practices are increas- ingly adopted in industrial settings, yet their integration with Opera- tional Technology (OT) systems presents significant challenges. This pa- per analyzes the fundamental obstacles in combining MLOps with OT en- vironments and proposes a systematic approach to embed MLOps prac- tices into established OT reference models. We evaluate the suitability of the Reference Architectural Model for Industry 4.0 (RAMI 4.0) and the International Society of Automation Standard 95 (ISA-95) for MLOps integration and present a detailed mapping of MLOps lifecycle compo- nents to RAMI 4.0 exemplified by a real-world use case. Our findings demonstrate that while standard MLOps practices cannot be directly transplanted to OT environments, structured adaptation using existing reference models can provide a pathway for successful integration.

cs.LG↗

arXiv2025

The Flood Complex: Large-Scale Persistent Homology on Millions of Points

We consider the problem of computing persistent homology (PH) for large-scale Euclidean point cloud data, aimed at downstream machine learning tasks, where the exponential growth of the most widely-used Vietoris-Rips complex imposes serious computational limitations. Although more scalable alternatives such as the Alpha complex or sparse Rips approximations exist, they often still result in a prohibitively large number of simplices. This poses challenges in the complex construction and in the subsequent PH computation, prohibiting their use on large-scale point clouds. To mitigate these issues, we introduce the Flood complex, inspired by the advantages of the Alpha and Witness complex constructions. Informally, at a given filtration value $r\geq 0$, the Flood complex contains all simplices from a Delaunay triangulation of a small subset of the point cloud $X$ that are fully covered by balls of radius $r$ emanating from $X$, a process we call flooding. Our construction allows for efficient PH computation, possesses several desirable theoretical properties, and is amenable to GPU parallelization. Scaling experiments on 3D point cloud data show that we can compute PH of up to dimension 2 on several millions of points. Importantly, when evaluating object classification performance on real-world and synthetic data, we provide evidence that this scaling capability is needed, especially if objects are geometrically or topologically complex, yielding performance superior to other PH-based methods and neural networks for point cloud data. Source code and datasets are available on https://github.com/plus-rkwitt/flooder.

cs.LG↗

arXiv2025

What are the limits to biomedical research acceleration through general-purpose AI?

Although general-purpose artificial intelligence (GPAI) is widely expected to accelerate scientific discovery, its practical limits in biomedicine remain unclear. We assess this potential by developing a framework of GPAI capabilities across the biomedical research lifecycle. Our scoping literature review indicates that current GPAI could deliver a speed increase of around 2x, whereas future GPAI could facilitate strong acceleration of up to 25x for physical tasks and 100x for cognitive tasks. However, achieving these gains may be severely limited by factors such as irreducible biological constraints, research infrastructure, data access, and the need for human oversight. Our expert elicitation with eight senior biomedical researchers revealed skepticism regarding the strong acceleration of tasks such as experiment design and execution. In contrast, strong acceleration of manuscript preparation, review and publication processes was deemed plausible. Notably, all experts identified the assimilation of new tools by the scientific community as a critical bottleneck. Realising the potential of GPAI will therefore require more than technological progress; it demands targeted investment in shared automation infrastructure and systemic reforms to research and publication practices.

cs.CY↗