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 a point cloud $X$ that are fully covered by the union of 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 GitHub: https://github.com/plus-rkwitt/flooder

This study conducts a comparative analysis of Model Predictive Control (MPC) and Proximal Policy Optimization (PPO), a Deep Reinforcement Learning (DRL) algorithm, applied to a 1-Degree of Freedom (DOF) Quanser Aero 2 system. Classical control techniques such as MPC and Linear Quadratic Regulator (LQR) are widely used due to their theoretical foundation and practical effectiveness. However, with advancements in computational techniques and machine learning, DRL approaches like PPO have gained traction in solving optimal control problems through environment interaction. This paper systematically evaluates the dynamic response characteristics of PPO and MPC, comparing their performance, computational resource consumption, and implementation complexity. Experimental results show that while LQR achieves the best steady-state accuracy, PPO excels in rise-time and adaptability, making it a promising approach for applications requiring rapid response and adaptability. Additionally, we have established a baseline for future RL-related research on this specific testbed. We also discuss the strengths and limitations of each control strategy, providing recommendations for selecting appropriate controllers for real-world scenarios.

We consider the problem of learning the dynamics in the topology of time-evolving point clouds, the prevalent spatiotemporal model for systems exhibiting collective behavior, such as swarms of insects and birds or particles in physics. In such systems, patterns emerge from (local) interactions among self-propelled entities. While several well-understood governing equations for motion and interaction exist, they are notoriously difficult to fit to data, as most prior work requires knowledge about individual motion trajectories, i.e., a requirement that is challenging to satisfy with an increasing number of entities. To evade such confounding factors, we investigate collective behavior from a _topological perspective_, but instead of summarizing entire observation sequences (as done previously), we propose learning a latent dynamical model from topological features _per time point_. The latter is then used to formulate a downstream regression task to predict the parametrization of some a priori specified governing equation. We implement this idea based on a latent ODE learned from vectorized (static) persistence diagrams and show that a combination of recent stability results for persistent homology justifies this modeling choice. Various (ablation) experiments not only demonstrate the relevance of each model component but provide compelling empirical evidence that our proposed model -- _Neural Persistence Dynamics_ -- substantially outperforms the state-of-the-art across a diverse set of parameter regression tasks.

Industry 4.0 factories are complex and data-driven. Data is yielded from many sources, including sensors, PLCs, and other devices, but also from IT, like ERP or CRM systems. We ask how to collect and process this data in a way, such that it includes metadata and can be used for industrial analytics or to derive intelligent support systems. This paper describes a new, query model based approach, which uses a big data architecture to capture data from various sources using OPC UA as a foundation. It buffers and preprocesses the information for the purpose of harmonizing and providing a holistic state space of a factory, as well as mappings to the current state of a production site. That information can be made available to multiple processing sinks, decoupled from the data sources, which enables them to work with the information without interfering with devices of the production, disturbing the network devices they are working in, or influencing the production process negatively. Metadata and connected semantic information is kept throughout the process, allowing to feed algorithms with meaningful data, so that it can be accessed in its entirety to perform time series analysis, machine learning or similar evaluations as well as replaying the data from the buffer for repeatable simulations.

Reinforcement Learning (RL) is a powerful machine learning paradigm that has been applied in various fields such as robotics, natural language processing and game playing achieving state-of-the-art results. Targeted to solve sequential decision making problems, it is by design able to learn from experience and therefore adapt to changing dynamic environments. These capabilities make it a prime candidate for controlling and optimizing complex processes in industry. The key to fully exploiting this potential is the seamless integration of RL into existing industrial systems. The industrial communication standard Open Platform Communications Unified Architecture (OPC UA) could bridge this gap.However, since RL and OPC UA are from different fields, there is a need for researchers to bridge the gap between the two technologies. This work serves to bridge this gap by providing a brief technical overview of both technologies and carrying out a semi-exhaustive literature review to gain insights on how RL and OPC UA are applied in combination.With this survey, three main research topics have been identified, following the intersection of RL with OPC UA. The results of the literature review show that RL is a promising technology for the control and optimization of industrial processes, but does not yet have the necessary standardized interfaces to be deployed in real-world scenarios with reasonably low effort.

The vision of Industry 4.0 introduces new requirements to Operational Technology (OT) systems. Solutions for these requirements already exist in the realm of Information Technology (IT), however, due to the different characteristics of IT and OT, these solutions often cannot be directly used in the realm of OT. We therefore propose an Industrial Business Process Twin (IBPT), allowing to apply methods of one world to another not directly but, instead, to a representation, that is in bidirectional exchange with the other world. The proposed IBPT entity acts as an intermediary, decoupling the realms of IT and OT, thus allowing for an integration of IT and OT components of different manufacturers and platforms. Using this approach, we demonstrate the four essential Industry 4.0 design principles information transparency, technical assistance, interconnection and decentralized decisions based on the gamified Industry 4.0 scenario of playing the game of Nine Men’s Morris. This scenario serves well for agent based Artificial Intelligence (AI)-research and education. We develop an Open Platform Communications Unified Architecture (OPC UA) information and communication model and then evaluate the IBPT component with respect to the different views of the Reference Architecture Model Industry 4.0 (RAMI4.0).

Topological data analysis offers a rich source of valuable information to study vision problems. Yet, so far we lack a theoretically sound connection to popular kernel-based learning techniques, such as kernel SVMs or kernel PCA. In this work, we establish such a connection by designing a multi-scale kernel for persistence diagrams, a stable summary representation of topological features in data. We show that this kernel is positive definite and prove its stability with respect to the 1-Wasserstein distance. Experiments on two benchmark datasets for 3D shape classification/retrieval and texture recognition show considerable performance gains of the proposed method compared to an alternative approach that is based on the recently introduced persistence landscapes.

We consider the problem of statistical computations with persistence diagrams, a summary representation of topological features in data. These diagrams encode persistent homology, a widely used invariant in topological data analysis. While several avenues towards a statistical treatment of the diagrams have been explored recently, we follow an alternative route that is motivated by the success of methods based on the embedding of probability measures into reproducing kernel Hilbert spaces. In fact, a positive definite kernel on persistence diagrams has recently been proposed, connecting persistent homology to popular kernel-based learning techniques such as support vector machines. However, important properties of that kernel which would enable a principled use in the context of probability measure embeddings remain to be explored. Our contribution is to close this gap by proving universality of a variant of the original kernel, and to demonstrate its effective use in two-sample hypothesis testing on synthetic as well as real-world data.

We investigate weighted straight skeletons from a geometric, graph-theoretical, and combinatorial point of view. We start with a thorough definition and shed light on some ambiguity issues in the procedural definition. We investigate the geometry, combinatorics, and topology of faces and the roof model, and we discuss in which cases a weighted straight skeleton is connected. Finally, we show that the weighted straight skeleton of even a simple polygon may be non-planar and may contain cycles, and we discuss under which restrictions on the weights and/or the input polygon the weighted straight skeleton still behaves similar to its unweighted counterpart. In particular, we obtain a non-procedural description and a linear-time construction algorithm for the straight skeleton of strictly convex polygons with arbitrary weights.

We investigate weighted straight skeletons from a geometric, graph-theoretical, and combinatorial point of view. We start with a thorough definition and shed light on some ambiguity issues in the procedural definition. We investigate the geometry, combinatorics, and topology of faces and the roof model, and we discuss in which cases a weighted straight skeleton is connected. Finally, we show that the weighted straight skeleton of even a simple polygon may be non-planar and may contain cycles, and we discuss under which restrictions on the weights and/or the input polygon the weighted straight skeleton still behaves similar to its unweighted counterpart. In particular, we obtain a non-procedural description and a linear-time construction algorithm for the straight skeleton of strictly convex polygons with arbitrary weights.

We study the characteristics of straight skeletons of monotone polygonal chains and use them to devise an algorithm for computing positively weighted straight skeletons of monotone polygons. Our algorithm runs in O(nlog⁡n) time and O(n) space, where n denotes the number of vertices of the polygon.

We study straight skeletons and make both theoretical and practical contributions which support new approaches to the computation of straight skeletons of arbitrary planar straight-line graphs (PSLGs). We start with an adequate extension of the concept of motorcycle graphs to PSLGs, with motorcycles starting at the reflex vertices of a PSLG, which allows us to generalize well-known results on the relation between the straight skeleton and the motorcycle graph to arbitrary PSLGs: the edges of the motorcycle graph cover a specific subset of the edges of the straight skeleton, and they form the basis of 3D slabs such that the projection of the lower envelope of those slabs to the plane forms the straight skeleton. As an immediate application we sketch how to use a graphics hardware for computing (approximate) straight skeletons of PSLGs. Further, we present and analyze a novel wavefront-type algorithm which bridges the current gap between the theory and practice of straight-skeleton computations. Our algorithm handles arbitrary PSLGs, is easy to implement, and is fast enough to handle complex data: it can be expected to run in O(n log n) time in practice for an n-vertex PSLG; its worst-case complexity is O(n2 log n). Extensive experimental results confirm an average runtime of 20 n log n µs on a standard PC for virtually all of our 13500 datasets of different characteristics. As also confirmed by our experiments, this constitutes an average gain in performance by a multiplicative factor of n, or at least one to two orders of magnitude, relative to the speed of the implementation provided by CGAL for closed polygons.

We study the watermarking of 2D vector data and introduce a framework which preserves topological properties of the input. Our framework is based on so-called maximum perturbation regions (MPR) of the input vertices, which is a concept similar to the just-noticeable-difference constraint. The MPRs are computed by means of the Voronoi diagram of the input and allow us to avoid (self-)intersections of input objects that might result from the embedding of the watermark. We demonstrate and analyze the applicability of this new framework by coupling it with a well-known approach to watermarking that is based on Fourier descriptors. However, our framework is general enough such that any robust scheme for the watermarking of vector data can be applied.