20005 research outputs found
Sort by
Enter Technico: a technically informed perspective on moral dilemmas in autonomous vehicle policy
No full textPEPbench - Open, reproducible, and systematic benchmarking of automated pre-ejection period extraction algorithms
No full textThe pre-ejection period (PEP) is a widely used cardiac parameter in psychophysiology that reflects the duration between the onset of ventricular depolarization and the opening of the aortic valve. PEP is often used as a marker of cardiac sympathetic nervous system (SNS) activity, particularly in within-subject comparisons under similar hemodynamic conditions. While many algorithms for automated PEP extraction from electrocardiography (ECG) and impedance cardiography (ICG) signals (more precisely, its first derivative, dZ/dt) have been proposed in literature, they have not been systematically benchmarked. This lack of standardized algorithm comparisons originates from the absence of open-source algorithms and annotated datasets for evaluating PEP extraction algorithms. To address this issue, we introduce PEPbench, an open-source Python package with different Q-peak and B-point detection algorithms from literature that can be combined to create comprehensive PEP extraction pipelines, and a standardized framework for evaluating PEP extraction algorithms. We use PEPbench to systematically compare 108 different algorithm combinations. All combinations are evaluated on two datasets with manually annotated Q-peaks and B-points, which we make publicly available as the first datasets with reference PEP annotations. Our results show that the algorithms can differ vastly in their performance and that B-point detection algorithms introduce a considerable amount of error. Thus, we suggest that automated PEP extraction algorithms should be used with caution on a beat-to-beat level as their error rates are relatively high. This highlights the need for open and reproducible benchmarking frameworks for PEP extraction algorithms to improve the quality of research findings in the field of psychophysiology. With PEPbench, we aim to take a first step toward this goal and encourage other researchers to engage in the evaluation of PEP extraction algorithms by contributing algorithms, data, and annotations. We hope to establish a community-driven platform, fostering innovation and collaboration in the field of psychophysiology and beyond
Toward variational quantum algorithms for generalized linear and nonlinear transport phenomena
No full textThis article proposes a variational quantum algorithm to solve linear and nonlinear thermofluid dynamic transport equations. The hybrid classical-quantum framework is applied to problems governed by the heat, wave, and Burgers’ equations in combination with different engineering boundary conditions. Topics covered include the encoding of band matrices, as in the consideration of nonconstant material properties and upwind-biased first- and higher-order approximations, widely used in engineering computational fluid dynamics, by the use of a mask function. Verification examples demonstrate high predictive agreement with classical methods. Furthermore, the scalability analysis shows a polylog scaling of the number of quantum gates with the number of qubits. Remaining challenges refer to the implicit construction of upwind schemes and the identification of an appropriate parameterization strategy of the quantum ansatz
Recycling of Ti-6Al-4V powder in metal binder jetting
No full textPowder in sinter-based metal binder jetting is not fused but only bonded, which in principle ensures full reusability of the unbonded powder. This increases both resource and cost efficiency. In order to explore this potential, this study investigates the effect of powder recycling with regard to handling and thermal influences. For this purpose, a gas atomized Ti-6Al-4V powder is used as a reference powder. It is shown that recycling the Ti-6Al-4V powder four times leads to a minor change in the particle size distribution and an improvement in flow properties. No differences in the average green part densities could be determined, but an increase in their range (from 1 % to up to 7 %). Furthermore, it is demonstrated that the thermal influence due to curing and conditioning (temperatures ≤ 200°C) after printing does not lead to a significant increase in oxygen or nitrogen even after 15 repetitions
Middle meningeal artery model for training endovascular treatment of chronic subdural hematomas in interventional neuroradiology
No full textTechnical progress and the development of smaller treatment instruments allow neurointerventional procedures to be used to treat diseases involving small vessels (< 2 mm). One example is the subdural hematoma (SDH), which can be treated by embolizing the middle meningeal artery (MMA) to cut off blood supply to SDH. The procedure was first used in 2018, following efficacy and safety studies. The embolization is technically very challenging and requires extensive training of the physicians. This work presents the development of an MMA model for endovascular training simulations of SDH with original instruments and particle embolization for integration into the existing neurointerventional training simulator Hamburg ANatomical NEurointerventional Simulator (HANNES). The development and testing were carried out by an interdisciplinary team of physicians and engineers. The aim of this work is to avoid the disadvantages of animal experiments, such as ethical aspects, anatomical differences to human vessel architecture, and long-term availability. First, a printing study of the MMA model was carried out to determine suitable processes and materials. Subsequently, suitable models were tested by experienced neurointerventional physicians in a realistic treatment setting, whereby they assessed 4 out of 20 models as sufficiently good. Relevant criteria were, among others, the flowrate, probing ability, elasticity, haptics, and geometric mapping. Based on these findings, an embolization module was developed to capture particles during training, which was evaluated as a moderate basic model for SDH embolization training. In conclusion, the novel MMA model with embolization module integrated in the simulator HANNES enables an innovative state-of-the-art neurointerventional training opportunity of physicians
Rolle der Digitalisierung für die Veränderungen in der Berufsbildung bis hin zur Neugestaltung von Lehr-Lern-Prozessen
No full textDie Digitalisierung der Lebens- und Arbeitswelt stellt die Berufsbildung und damit das Lehren und Lernen vor völlig neue Herausforderungen. Die Konsequenzen für die Berufsbildung reichen von der Frage „Sind die Berufe noch zeitgemäß?" bis hin zur Frage „Muss das Lernen völlig neu gedacht werden?". Der Beitrag soll aufzeigen, wo die Chancen und Herausforderungen für Veränderungen in den Lehr- und Lernprozessen in der beruflichen Bildung liegen
The influence of AI support on product development results: a live-lab case study of a top-heat grill
No full textThis study examines the influence of AI tools on development speed, functionality, costs and team satisfaction in a university design project. Two engineering teams developed solutions for a top-heat grill within three development sprints. The test group used AI tools from the start, the control group from the third sprint onwards. The comparison, based on interviews and the evaluation from the software tools used, shows that the use of AI tools has a positive impact on development speed, functionality and costs. These findings support the understanding of the potential of AI tools in development projects
Foresee: ML-driven, communication-efficient time-series forecasting
No full textIn the Internet of Things, a multitude of sensors continuously collect data and transmit it to the cloud for analysis. However, frequent transfer of measurements is impractical for battery-powered sensors due to the high energy-costs of wireless communication. Therefore, sensors often collect data and send it at periodic intervals, while a state-of-the-art cloud-based predictive model estimates intermediate values between transmissions. This paper introduces Foresee, which improves data quality on the cloud without additional communication overhead compared to periodic and model-drives approaches. Foresee makes local predictions on the sensor by employing a small, resource-efficient neural network. Upon detecting significant deviations between predicted and measured data, Foresee communicates these measurements to the cloud. Thus, Foresee notifies the cloud whenever predictions are difficult. On the cloud side, Foresee uses a state-of-the-art transformer model to make predictions between transmissions. Our results demonstrate the effectiveness of Foresee across different datasets. For instance, on the AlSolar dataset, with a prediction length of 48, we observe a 26% improvement in the Mean Absolute Error without any additional communication, compared to periodic communication every 39 timesteps. Additionally, Foresee achieves a 63% reduction in Mean Absolute Error compared to a model-driven approach with the same communication overhead
Industrialization of additive manufacturing
No full textOur world is constantly facing a wide variety of social challenges – currently, issues such as climate change, security, demographic change, and resource conservation are of significant importance and show an acute need for action. Disruptive technologies, among other things, are needed to tackle such problems. Additive manufacturing (AM) is regarded as the ‘game changer / paradigm shift’ technology for the digital, automated production of the future – as the key enabler of ‘Production 2.0’
Discrete prompt optimization using genetic algorithm for secure Python code generation
No full textLarge language models (LLMs) have become powerful tools that enable novice developers to generate production-level code. However, research has highlighted the security risks associated with such code generation, due to the high volume of generated software vulnerabilities. Recent studies have explored various techniques for automatically optimizing prompts to elicit desired responses from LLMs. Among these methods, Genetic Algorithms (GAs), which search for optimal solutions by evolving an initial population of candidates through iterative mutations, have gained attention as a lightweight and effective prompt optimization approach that does not require large datasets or access to model weights. However, their potential has not yet been examined in the context of secure code generation. In this paper, we use GA to develop a discrete prompt optimization pipeline specifically designed for secure code generation. We introduce two domain-specific prompt mutation techniques and assess how incorporating these security-focused mutations alongside general-purpose techniques, such as back translation and paraphrasing, affects the security of Python code generated by LLMs. Results demonstrate that our security-specific mutation techniques led to prompts with richer security context compared to the generic mutation techniques. Furthermore, combining these techniques with generic mutations substantially reduced the number of security weaknesses in the LLM-generated code. We also observed that prompts optimized for a particular LLM tend to perform best on that same model, highlighting the importance of model-specific prompt optimization