Search

We compute how elastic deformations of optical fibers affect light propagation therein. Specifically, we consider differences in wave-guiding properties of straight fibers subject to different external temperatures, pressures, and gravitational fields. This is done by solving, perturbatively to first order, the Maxwell equations in deformed and anisotropic fibers using a multiple-scales approximation scheme. We derive explicit expressions for the induced phase shift and birefringence. The phase shift can be expressed in terms of the average radial pressure, longitudinal tension, and change in temperature, while birefringence depends on the quadrupole of the external pressure distribution and the stresses on the axis of the fiber.

The Burrows-Wheeler transform (BWT) is integral to the FM-index, which is used extensively in text compression, indexing, pattern search, and bioinformatic problems as de novo assembly and read alignment. Thus, efficient construction of the BWT in terms of time and memory usage is key to these applications. We present a novel external algorithm called Improved-Bucket Burrows-Wheeler transform (IBB) for constructing the BWT of DNA datasets with highly diverse sequence lengths. IBB uses a right-aligned approach to efficiently handle sequences of varying lengths, a tree-based data structure to manage relative insert positions and ranks, and fine buckets to reduce the necessary amount of input and output to external memory. Our experiments demonstrate that IBB is 10% to 40% faster than the best existing state-of-the-art BWT construction algorithms on most datasets while maintaining competitive memory consumption.

Timber represents an increasingly valuable and versatile resource. However, forestry operations such as harvesting, handling and measuring logs still require substantial human labor in remote environments posing significant safety risks. Progressively automating these tasks has the potential of increasing their efficiency as well as safety, but requires an accurate detection of individual logs as well as live trees and their context. Although initial approaches have been proposed for this challenging application domain, specialized data and algorithms are still too scarce to develop robust solutions. To mitigate this gap, we introduce the TimberVision dataset, consisting of more than 2k annotated RGB images containing a total of 51k trunk components including cut and lateral surfaces, thereby surpassing any existing dataset in this domain in terms of both quantity and detail by a large margin. Based on this data, we conduct a series of ablation experiments for oriented object detection and instance segmentation and evaluate the influence of multiple scene parameters on model performance. We introduce a generic framework to fuse the components detected by our models for both tasks into unified trunk representations. Furthermore, we automatically derive geometric properties and apply multi-object tracking to further enhance robustness. Our detection and tracking approach provides highly descriptive and accurate trunk representations solely from RGB image data, even under challenging environmental conditions. Our solution is suitable for a wide range of application scenarios and can be readily combined with other sensor modalities., Comment: Accepted at Winter Conference on Applications of Computer Vision (WACV) 2025. Code and dataset available at https://github.com/timbervision/timbervision

This paper introduces a recurrent neural network approach for predicting user lifetime value in Software as a Service (SaaS) applications. The approach accounts for three connected time dimensions. These dimensions are the user cohort (the date the user joined), user age-in-system (the time since the user joined the service) and the calendar date the user is an age-in-system (i.e., contemporaneous information).The recurrent neural networks use a multi-cell architecture, where each cell resembles a long short-term memory neural network. The approach is applied to predicting both acquisition (new users) and rolling (existing user) lifetime values for a variety of time horizons. It is found to significantly improve median absolute percent error versus light gradient boost models and Buy Until You Die models., Comment: 4 pages, 4 figures, first presented in Customer Journey Optimization Workshop in KDD 2022

Computer-Assisted Interventions enable clinicians to perform precise, minimally invasive procedures, often relying on advanced imaging methods. Cone-beam computed tomography (CBCT) can be used to facilitate computer-assisted interventions, despite often suffering from artifacts that pose challenges for accurate interpretation. While the degraded image quality can affect image analysis, the availability of high quality, preoperative scans offers potential for improvements. Here we consider a setting where preoperative CT and intraoperative CBCT scans are available, however, the alignment (registration) between the scans is imperfect to simulate a real world scenario. We propose a multimodal learning method that fuses roughly aligned CBCT and CT scans and investigate the effect on segmentation performance. For this experiment we use synthetically generated data containing real CT and synthetic CBCT volumes with corresponding voxel annotations. We show that this fusion setup improves segmentation performance in $18$ out of $20$ investigated setups., Comment: Accepted at BVM 2025. arXiv admin note: text overlap with arXiv:2406.11650

Researching novel user experiences in medicine is challenging due to limited access to equipment and strict ethical protocols. Extended Reality (XR) simulation technologies offer a cost- and time-efficient solution for developing interactive systems. Recent work has shown Extended Reality Prototyping (XRP)'s potential, but its applicability to specific domains like controlling complex machinery needs further exploration. This paper explores the benefits and limitations of XRP in controlling a mobile medical imaging robot. We compare two XR visualization techniques to reduce perceived latency between user input and robot activation. Our XRP validation study demonstrates its potential for comparative studies, but identifies a gap in modeling human behavior in the analytic XRP validation framework.

The RD50-MPW4 is the latest HV-CMOS pixel sensor from the CERN-RD50-CMOS working group, designed to evaluate the HV-CMOS technology in terms of spatial resolution, radiation hardness and timing performance. Fabricated by LFoundry using a 150nm process, it features an improved architecture to mitigate crosstalk, which has been an issue with the predecessor RD50-MPW3, allowing more sensitive threshold settings and full matrix operation. Enhancements include separated power domains for peripheral and in-pixel digital readout, a new backside-biasing step, and an improved guard ring structure supporting biasing up to 500V, significantly boosting radiation hardness. Laboratory measurements and test beam results presented in this paper show significant improvements over its predecessor regarding noise behavior, spatial resolution, and efficiency., Comment: Preprint version of Proceedings of Pisa meeting 2024

Medical imaging is vital in computer assisted intervention. Particularly cone beam computed tomography (CBCT) with defacto real time and mobility capabilities plays an important role. However, CBCT images often suffer from artifacts, which pose challenges for accurate interpretation, motivating research in advanced algorithms for more effective use in clinical practice. In this work we present CBCTLiTS, a synthetically generated, labelled CBCT dataset for segmentation with paired and aligned, high quality computed tomography data. The CBCT data is provided in 5 different levels of quality, reaching from a large number of projections with high visual quality and mild artifacts to a small number of projections with severe artifacts. This allows thorough investigations with the quality as a degree of freedom. We also provide baselines for several possible research scenarios like uni- and multimodal segmentation, multitask learning and style transfer followed by segmentation of relatively simple, liver to complex liver tumor segmentation. CBCTLiTS is accesssible via https://www.kaggle.com/datasets/maximiliantschuchnig/cbct-liver-and-liver-tumor-segmentation-train-data., Comment: Accepted at VCBM 2024 - https://conferences.eg.org/vcbm2024/

Gravitational wave memory is said to arise when a gravitational wave burst produces changes in a physical system that persist even after that wave has passed. This paper analyzes gravitational wave bursts in plane wave spacetimes, deriving memory effects for timelike and null geodesics, massless scalar fields, and massless spinning particles whose motion is described by the spin Hall equations. We find that all such effects are characterized by four "memory tensors," three of which are independent. We also show that memory effects for null geodesics can have strong longitudinal components, even in vacuum general relativity. When considering massless particles with spin, we solve the spin Hall equations analytically by showing that there exists a conservation law associated with each conformal Killing vector. For the scattering of fields by gravitational waves, we show that given any solution to the massless scalar field equation in flat spacetime, a weak-field solution in a plane wave spacetime can be generated just by applying an appropriate differential operator -- an operator that is constructed from the aforementioned memory tensors. Memory effects for scalar fields are illustrated for both incoming plane waves and higher-order Gaussian beams. We also present a numerical comparison between the spin Hall equations and the full evolution of localized wave packets with angular momentum. Although we work in plane wave spacetimes, which are physically idealized, similar results are also expected to apply for sufficiently small systems affected by distantly generated gravitational waves. Using the Penrose limit, our results may also apply to ultrarelativistic systems at arbitrary locations in arbitrary (even nonradiating) spacetimes.

Cone-beam computed tomography (CBCT) is an important tool facilitating computer aided interventions, despite often suffering from artifacts that pose challenges for accurate interpretation. While the degraded image quality can affect downstream segmentation, the availability of high quality, preoperative scans represents potential for improvements. Here we consider a setting where preoperative CT and intraoperative CBCT scans are available, however, the alignment (registration) between the scans is imperfect. We propose a multimodal learning method that fuses roughly aligned CBCT and CT scans and investigate the effect of CBCT quality and misalignment on the final segmentation performance. For that purpose, we make use of a synthetically generated data set containing real CT and synthetic CBCT volumes. As an application scenario, we focus on liver and liver tumor segmentation. We show that the fusion of preoperative CT and simulated, intraoperative CBCT mostly improves segmentation performance (compared to using intraoperative CBCT only) and that even clearly misaligned preoperative data has the potential to improve segmentation performance., Comment: Submitted to SASHIMI2024 (MICCAI workshop)

We analyse the deformations of a cylindrical elastic body resulting from displacements in a varying gravitational field., Comment: 31 pages, 21 figures

Artificial Intelligence (AI) systems, trained in controlled environments, often struggle in real-world complexities. We propose a general framework for estimating domain complexity across diverse environments, like open-world learning and real-world applications. This framework distinguishes between intrinsic complexity (inherent to the domain) and extrinsic complexity (dependent on the AI agent). By analyzing dimensionality, sparsity, and diversity within these categories, we offer a comprehensive view of domain challenges. This approach enables quantitative predictions of AI difficulty during environment transitions, avoids bias in novel situations, and helps navigate the vast search spaces of open-world domains., Comment: 34 pages, 13 figures, 7 tables. arXiv admin note: substantial text overlap with arXiv:2303.04141

Semantic segmentation is a crucial task in medical image processing, essential for segmenting organs or lesions such as tumors. In this study we aim to improve automated segmentation in CBCTs through multi-task learning. To evaluate effects on different volume qualities, a CBCT dataset is synthesised from the CT Liver Tumor Segmentation Benchmark (LiTS) dataset. To improve segmentation, two approaches are investigated. First, we perform multi-task learning to add morphology based regularization through a volume reconstruction task. Second, we use this reconstruction task to reconstruct the best quality CBCT (most similar to the original CT), facilitating denoising effects. We explore both holistic and patch-based approaches. Our findings reveal that, especially using a patch-based approach, multi-task learning improves segmentation in most cases and that these results can further be improved by our denoising approach., Comment: Accepted and presented at German Conference on Medical Image Computing (BVM) 2024 edit: During work on this publication Maximilian Ernst Tschuchnig was affiliated with Salzburg University of Applied Sciences and University of Salzburg

Results in this paper present an in-depth study of time resolution for active pixels of the RD50-MPW2 prototype CMOS particle detector. Measurement techniques employed include Backside- and Edge-TCT configurations, in addition to electrons from a $^{90}\mathrm{Sr}$ source. A sample irradiated to $5\cdot 10^{14}\,\mathrm{n}_\mathrm{eq}/\mathrm{cm}^2$ was used to study the effect of radiation damage. Timing performance was evaluated for the entire pixel matrix and with positional sensitivity within individual pixels as a function of the deposited charge. Time resolution obtained with TCT is seen to be uniform throughout the pixel's central region with approx. $220\,\mathrm{ps}$ at $12\,\mathrm{ke}^-$ of deposited charge, degrading at the edges and lower values of deposited charge. $^{90}\mathrm{Sr}$ measurements show a slightly worse time resolution as a result of delayed events coming from the peripheral areas of the pixel.

We analyze the Cauchy problem for the Proca equation in gravitating dielectric media., Comment: 6 pages

By design, quasi delay-insensitive (QDI) circuits exhibit higher resilience against timing variations as compared to their synchronous counterparts. Since computation in QDI circuits is event-based rather than clock-triggered, spurious events due to transient faults such as radiation-induced glitches, a priori are of higher concern in QDI circuits. In this work we propose a formal framework with the goal to gain a deeper understanding on how susceptible QDI circuits are to transient faults. We introduce a worst-case model for transients in circuits. We then prove an equivalence of faults within this framework and use this result to provably exhaustively check QDI circuits, a linear Muller pipeline and a cyclic Muller pipeline, for their susceptibility to produce non-stable output signals.

Deep Learning (DL) systems have proliferated in many applications, requiring specialized hardware accelerators and chips. In the nano-era, devices have become increasingly more susceptible to permanent and transient faults. Therefore, we need an efficient methodology for analyzing the resilience of advanced DL systems against such faults, and understand how the faults in neural accelerator chips manifest as errors at the DL application level, where faults can lead to undetectable and unrecoverable errors. Using fault injection, we can perform resilience investigations of the DL system by modifying neuron weights and outputs at the software-level, as if the hardware had been affected by a transient fault. Existing fault models reduce the search space, allowing faster analysis, but requiring a-priori knowledge on the model, and not allowing further analysis of the filtered-out search space. Therefore, we propose ISimDL, a novel methodology that employs neuron sensitivity to generate importance sampling-based fault-scenarios. Without any a-priori knowledge of the model-under-test, ISimDL provides an equivalent reduction of the search space as existing works, while allowing long simulations to cover all the possible faults, improving on existing model requirements. Our experiments show that the importance sampling provides up to 15x higher precision in selecting critical faults than the random uniform sampling, reaching such precision in less than 100 faults. Additionally, we showcase another practical use-case for importance sampling for reliable DNN design, namely Fault Aware Training (FAT). By using ISimDL to select the faults leading to errors, we can insert the faults during the DNN training process to harden the DNN against such faults. Using importance sampling in FAT reduces the overhead required for finding faults that lead to a predetermined drop in accuracy by more than 12x., Comment: Submission under review

Dynamic partial reconfiguration enables multi-tenancy in cloud-based FPGAs, which presents security challenges for tenants, IPs, and data. Malicious users can exploit FPGAs for remote side-channel attacks (SCAs), and shared on-chip resources can be used for attacks. Logical separation can ensure design integrity, but on-chip resources can still be exploited. Conventional SCA mitigation can help, but it requires significant effort, and bitstream checking techniques are not highly accurate. An active on-chip defense mechanism is needed for tenant confidentiality. Toward this, we propose a lightweight shielding technique utilizing ring oscillators (ROs) to protect applications against remote power SCA. Unlike existing RO-based approaches, in our methodology, an offline pre-processing stage is proposed to carefully configure power monitors and an obfuscating circuit concerning the resource constraints of the board. Detection of power fluctuations due to application execution enables the obfuscating circuit to flatten the power consumption trace. To evaluate the effectiveness of the proposed SHIELD, we implemented it on a Xilinx Zynq-7000 FPGA board executing an RSA encryption algorithm. Due to the SHIELD, the number of traces required to extract the encryption key is increased by 166x, making an attack extremely hard at run-time. Note that the proposed SHIELD does not require any modification in the target application. Our methodology also shows up to 54% less power consumption and up to 26% less area overhead than the state-of-the-art random noise-addition-based defense., Comment: 11 pages, under review by IEEE ACCESS

The topic of a photon mass has received justified attention over the years with strong theoretical reasoning for its existence. We derive a full description of mode solutions in step-index optical fibres on the basis of Proca's equation as a model for massive photons. The no-go theorem recently found in coaxial waveguides does not appear in the optical fibre case., Comment: 6 pages, 1 figure

As artificial intelligence research advances, the platforms used to evaluate AI agents need to adapt and grow to continue to challenge them. We present the Polycraft World AI Lab (PAL), a task simulator with an API based on the Minecraft mod Polycraft World. Our platform is built to allow AI agents with different architectures to easily interact with the Minecraft world, train and be evaluated in multiple tasks. PAL enables the creation of tasks in a flexible manner as well as having the capability to manipulate any aspect of the task during an evaluation. All actions taken by AI agents and external actors (non-player-characters, NPCs) in the open-world environment are logged to streamline evaluation. Here we present two custom tasks on the PAL platform, one focused on multi-step planning and one focused on navigation, and evaluations of agents solving them. In summary, we report a versatile and extensible AI evaluation platform with a low barrier to entry for AI researchers to utilize., Comment: 27 pages, 5 figures

We show that the Proca equation in vacuum, as well as its plausible modifications in dielectric media, is incompatible with experimental evidence, no matter how small the Proca mass is., Comment: 6 pages; minor changes to formatting

Bluetooth (BT) has revolutionized close-range communication enabling smart capabilities in everyday devices through wireless technology. One of the most important sub-domains of Internet-of-Things (IoT) specializes in the usage of BT technologies to develop smart homes and environments, which include hospitals, buildings, shopping facilities, etc. to offer a wide-range of features, like instantaneous and remote access to ventilation, lighting, security, localization, and tracking. However, the deployment of such features in smart infrastructures are typically unaccompanied by appropriate security measures that safeguard the data and protect its users. Towards this, we propose the ATLAS framework, which is composed of our novel IoT architecture and secure networking stack that can be used to anonymously localize and track smartphones and wearables by deploying multiple Bluetooth Low Energy (BLE) beacons across the environment. The proposed networking stack enables varying levels of encryption across all layers of the communication stack to ensure an easy-to-adopt, secure-by-design network architecture. We also deploy a novel data transformation and fingerprinting-based localization algorithm, which is highly effective in localizing user devices within a given area. The ATLAS framework is open-sourced at https://atlas-tuw.sourceforge.io to enable wide-spread adoption and further research and development.

Common Deep Metric Learning (DML) datasets specify only one notion of similarity, e.g., two images in the Cars196 dataset are deemed similar if they show the same car model. We argue that depending on the application, users of image retrieval systems have different and changing similarity notions that should be incorporated as easily as possible. Therefore, we present Language-Guided Zero-Shot Deep Metric Learning (LanZ-DML) as a new DML setting in which users control the properties that should be important for image representations without training data by only using natural language. To this end, we propose InDiReCT (Image representations using Dimensionality Reduction on CLIP embedded Texts), a model for LanZ-DML on images that exclusively uses a few text prompts for training. InDiReCT utilizes CLIP as a fixed feature extractor for images and texts and transfers the variation in text prompt embeddings to the image embedding space. Extensive experiments on five datasets and overall thirteen similarity notions show that, despite not seeing any images during training, InDiReCT performs better than strong baselines and approaches the performance of fully-supervised models. An analysis reveals that InDiReCT learns to focus on regions of the image that correlate with the desired similarity notion, which makes it a fast to train and easy to use method to create custom embedding spaces only using natural language., Comment: Accepted to WACV 2023

Research on Deep Neural Networks (DNNs) has focused on improving performance and accuracy for real-world deployments, leading to new models, such as Spiking Neural Networks (SNNs), and optimization techniques, e.g., quantization and pruning for compressed networks. However, the deployment of these innovative models and optimization techniques introduces possible reliability issues, which is a pillar for DNNs to be widely used in safety-critical applications, e.g., autonomous driving. Moreover, scaling technology nodes have the associated risk of multiple faults happening at the same time, a possibility not addressed in state-of-the-art resiliency analyses. Towards better reliability analysis for DNNs, we present enpheeph, a Fault Injection Framework for Spiking and Compressed DNNs. The enpheeph framework enables optimized execution on specialized hardware devices, e.g., GPUs, while providing complete customizability to investigate different fault models, emulating various reliability constraints and use-cases. Hence, the faults can be executed on SNNs as well as compressed networks with minimal-to-none modifications to the underlying code, a feat that is not achievable by other state-of-the-art tools. To evaluate our enpheeph framework, we analyze the resiliency of different DNN and SNN models, with different compression techniques. By injecting a random and increasing number of faults, we show that DNNs can show a reduction in accuracy with a fault rate as low as 7 x 10 ^ (-7) faults per parameter, with an accuracy drop higher than 40%. Run-time overhead when executing enpheeph is less than 20% of the baseline execution time when executing 100 000 faults concurrently, at least 10x lower than state-of-the-art frameworks, making enpheeph future-proof for complex fault injection scenarios. We release enpheeph at https://github.com/Alexei95/enpheeph., Comment: Source code: https://github.com/Alexei95/enpheeph To appear at 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), October, 2022

Time series are ubiquitous and therefore inherently hard to analyze and ultimately to label or cluster. With the rise of the Internet of Things (IoT) and its smart devices, data is collected in large amounts any given second. The collected data is rich in information, as one can detect accidents (e.g. cars) in real time, or assess injury/sickness over a given time span (e.g. health devices). Due to its chaotic nature and massive amounts of datapoints, timeseries are hard to label manually. Furthermore new classes within the data could emerge over time (contrary to e.g. handwritten digits), which would require relabeling the data. In this paper we present SuSL4TS, a deep generative Gaussian mixture model for semi-unsupervised learning, to classify time series data. With our approach we can alleviate manual labeling steps, since we can detect sparsely labeled classes (semi-supervised) and identify emerging classes hidden in the data (unsupervised). We demonstrate the efficacy of our approach with established time series classification datasets from different domains.

Recent studies have shown that many deep metric learning loss functions perform very similarly under the same experimental conditions. One potential reason for this unexpected result is that all losses let the network focus on similar image regions or properties. In this paper, we investigate this by conducting a two-step analysis to extract and compare the learned visual features of the same model architecture trained with different loss functions: First, we compare the learned features on the pixel level by correlating saliency maps of the same input images. Second, we compare the clustering of embeddings for several image properties, e.g. object color or illumination. To provide independent control over these properties, photo-realistic 3D car renders similar to images in the Cars196 dataset are generated. In our analysis, we compare 14 pretrained models from a recent study and find that, even though all models perform similarly, different loss functions can guide the model to learn different features. We especially find differences between classification and ranking based losses. Our analysis also shows that some seemingly irrelevant properties can have significant influence on the resulting embedding. We encourage researchers from the deep metric learning community to use our methods to get insights into the features learned by their proposed methods., Comment: Published at ICCV 2021

Electrification of all economic sectors and solar photovoltaics (PV) becoming the lowest-cost electricity generation technology in ever more regions give rise to new potential gains of trade. We develop a stylized analytical model to minimize unit energy cost in autarky, open it to different trade configurations, and evaluate it empirically. We identify large potential gains from interhemispheric and global electricity trade by combining complementary seasonal and diurnal cycles. The corresponding high willingness to pay for large-scale transmission suggests far-reaching political economy and regulatory implications.

The RD50-CMOS group aims to design and study High Voltage CMOS (HVCMOS) chips for use in a high radiation environment. Currently, measurements are performed on RD50-MPW2 chip, the second prototype developed by this group. The active matrix of the prototype consists of 8x8 pixels with analog front end. Details of the analog front end and simulations have been already published earlier. This contribution focuses on the Caribou based readout system of the active matrix. Each pixel of the active matrix can be readout one after the other. Relevant aspects of hardware, firmware and software are introduced. As a first stage, firmware for a standalone setup is introduced and details on data flow are given. Afterwards, a second stage of the firmware capable of synchronizing with other detectors and accepting triggers is presented, focusing on operation of the chip in combination with a tracking telescope to measure efficiency and residuals., Comment: Published under licence in Journal of Physics: Conference Series by IOP Publishing Ltd. CC-BY Content from this work may be used under the terms of the Creative Commons Attribution 4.0 International licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI

A polyhedron $\textbf{P} \subset \mathbb{R}^3$ has Rupert's property if a hole can be cut into it, such that a copy of $\textbf{P}$ can pass through this hole. There are several works investigating this property for some specific polyhedra: for example, it is known that all 5 Platonic and 9 out of the 13 Archimedean solids admit Rupert's property. A commonly believed conjecture states that every convex polyhedron is Rupert. We prove that Rupert's problem is algorithmically decidable for polyhedra with algebraic coordinates. We also design a probabilistic algorithm which can efficiently prove that a given polyhedron is Rupert. Using this algorithm we not only confirm this property for the known Platonic and Archimedean solids, but also prove it for one of the remaining Archimedean polyhedra and many others. Moreover, we significantly improve on almost all known Nieuwland numbers and finally conjecture, based on statistical evidence, that the Rhombicosidodecahedron is in fact not Rupert., Comment: To appear in Mathematics of Computation