JUGGLER: autonomous cost optimization and performance prediction of big data applications. - In: The ACM digital library, (2022), S. 1840-1854
Distributed in-memory processing frameworks accelerate iterative workloads by caching suitable datasets in memory rather than recomputing them in each iteration. Selecting appropriate datasets to cache as well as allocating a suitable cluster configuration for caching these datasets play a crucial role in achieving optimal performance. In practice, both are tedious, time-consuming tasks and are often neglected by end users, who are typically not aware of workload semantics, sizes of intermediate data, and cluster specification. To address these problems, we present Juggler, an end-to-end framework, which autonomously selects appropriate datasets for caching and recommends a correspondingly suitable cluster configuration to end users, with the aim of achieving optimal execution time and cost. We evaluate Juggler on various iterative, real-world, machine learning applications. Compared with our baseline, Juggler reduces execution time to 25.1% and cost to 58.1%, on average, as a result of selecting suitable datasets for caching. It recommends optimal cluster configuration in 50% of cases and near-to-optimal configuration in the remaining cases. Moreover, Juggler achieves an average performance prediction accuracy of 90%.
Joint model order estimation for multiple tensors with a coupled mode and applications to the joint decomposition of EEG, MEG Magnetometer, and Gradiometer tensors. - In: IEEE Xplore digital library, ISSN 2473-2001, (2022), S. 1186-1190
The efficient estimation of an approximate model order is essential for applications with multidimensional data if the observed low-rank data is corrupted by additive noise. Certain signal processing applications such as biomedical studies, where the data are collected simultaneously through heterogeneous sensors, share some common features, i.e., coupled factors among multiple tensors. The exploitation of this coupling can lead to a better model order estimation, especially in case of low SNRs. In this paper, we extend the rank estimation techniques, designed for a single tensor, to noise-corrupted coupled low-rank tensors that share one of their factor matrices. To this end, we consider the joint effect of the global eigenvalues (calculated from the coupled HOSVD) and exploit the exponential behavior of the resulting coupled global eigenvalues. We show that the proposed method outperforms the classical criteria and can be successfully applied to EEG, MEG Magnetometer, and Gradiometer measurements. Our real data simulation results show that the estimated rank is highly reliable in terms of dominant components extraction.
Cognitive network function for mobility robustness optimization in cellular networks. - In: IEEE Xplore digital library, ISSN 2473-2001, (2022), S. 2025-2040
Self Organizing Networks (SON) aim at automating different network management functions, thereby improving their efficiency while reducing the operational expenditures. There are several proposed SON Functions (SFs) in the standards and a crucial one among them is Mobility Robustness Optimization (MRO). It focuses on providing seamless connectivity to mobile User Equipments (UEs). While optimizing handovers, there is a trade off between the Radio Link Failures (RLFs) and ping-pongs. Research has been widely done on the applicability of machine learning algorithms in SON for making decisions in a cognitive manner. In this study, MRO problem is modeled in two ways using two different classes of machine learning algorithms - Regression (linear and non-linear) and Recommender System. The work is evaluated on a Long Term Evolution (LTE) network simulator for different traffic scenarios. It is observed that the recommender system based solution has an edge over the regression based approaches and there is an overall improvement of 3.7% in the handover performance compared to that of the baseline approach.
Open sub-granting radio resources in overlay D2D-based V2V communications. - In: EURASIP journal on wireless communications and networking, ISSN 1687-1499, Bd. 2022 (2022), 46, S. 1-29
Richtiger Name des Verfassers: Dariush Mohammad Soleymani
Capacity, reliability, and latency are seen as key requirements of new emerging applications, namely vehicle-to-everything (V2X) and machine-type communication in future cellular networks. D2D communication is envisaged to be the enabler to accomplish the requirements for the applications as mentioned earlier. Due to the scarcity of radio resources, a hierarchical radio resource allocation, namely the sub-granting scheme, has been considered for the overlay D2D communication. In this paper, we investigate the assignment of underutilized radio resources from D2D communication to device-to-infrastructure communication, which are moving in a dynamic environment. The sub-granting assignment problem is cast as a maximization problem of the uplink cell throughput. Firstly, we evaluate the sub-granting signaling overhead due to mobility in a centralized sub-granting resource algorithm, dedicated sub-granting radio resource (DSGRR), and then a distributed heuristics algorithm, open sub-granting radio resource (OSGRR), is proposed and compared with the DSGRR algorithm and no sub-granting case. Simulation results show improved cell throughput for the OSGRR compared with other algorithms. Besides, it is observed that the overhead incurred by the OSGRR is less than the DSGRR while the achieved cell throughput is yet close to the maximum achievable uplink cell throughput.
Expanding the remote experiment set with the 3Axis Portal physical model. - In: International journal of online and biomedical engineering, ISSN 2626-8493, Bd. 18 (2022), 04, S. 21-30
The problem of insufficient variety of experiments with physical models in laboratory workshops for distance learning in the design of control systems is posed. It is caused by the limited operations with the physical model when using the interface at the level of electromechanics control. The way of solving the problem is substantiated: the transition from electromechanics control of physical models of devices to control systems for the processes of using these devices. The proposed way to increase the number of types of experiments is illustrated with examples of systems for using popular physical models of an elevator, storage warehouse and production cell. However, the main focus is on expanding the functionality of the physical model 3-Axis Portal. These possibilities are realized by equipping the portal head of the base model with new sensors and actuators, improving the loads with which the portal works and using new types of active surfaces in the working field of the portal. Based on the aggregation of the proposed nodes with the basic portal model, sixteen types of systems for setting up remote experiments are described. The structures of these systems, elements of implementation and variants of experiments are described, which relate to the design of digital control systems, visualization of sorting algorithms, technical diagnostics of electronic assemblies, pattern recognition and other relevant topics for teaching students of engineering specialties.
Estimating food ingredient compositions based on mandatory product labeling. - In: Journal of food composition and analysis, ISSN 0889-1575, Bd. 110 (2022), 104508, S. 1-9
Having a specific understanding of the actual ingredient composition of products helps to calculate additional nutritional information, such as containing fatty and amino acids, minerals and vitamins, as well as to determine its environmental impacts. Unfortunately, producers rarely provide information on how much of each ingredient is in a product. Food manufacturers are, however, required to declare their products in terms of a label comprising an ingredient list (in descending order) and Big7 nutrient values. In this paper, we propose an automated approach for estimating ingredient contents in food products. First, we parse product labels to extract declared ingredients. Next, we exert mathematical formulations on the assumption that the weighted sum of Big7 ingredients as available from food compositional tables should resemble the product’s declared overall Big7 composition. We apply mathematical optimization techniques to find the best fitting ingredient composition estimate. We apply the proposed method to a dataset of 1804 food products spanning 11 product categories. We find that 76% of these products could be analyzed by our approach, and a composition within the prescribed nutrient tolerances could be calculated, using 20% of the allowed tolerances per Big7 ingredient on average. The remaining 24% of the food products could still be estimated when relaxing one or multiple nutrient tolerances. A study with known ingredient compositions shows that estimates are within a 0.9% difference of products’ actual recipes. Hence, the automated approach presented here allows for further analysis of large product quantities and provides possibilities for more intensive nutritional and ecological evaluations of food.
Individually optimized multi-channel tDCS for targeting somatosensory cortex. - In: Clinical neurophysiology, ISSN 1872-8952, Bd. 134 (2022), S. 9-26
Objective - Transcranial direct current stimulation (tDCS) is a non-invasive neuro-modulation technique that delivers current through the scalp by a pair of patch electrodes (2-Patch). This study proposes a new multi-channel tDCS (mc-tDCS) optimization method, the distributed constrained maximum intensity (D-CMI) approach. For targeting the P20/N20 somatosensory source at Brodmann area 3b, an integrated combined magnetoencephalography (MEG) and electroencephalography (EEG) source analysis is used with individualized skull conductivity calibrated realistic head modeling. - Methods - Simulated electric fields (EF) for our new D-CMI method and the already known maximum intensity (MI), alternating direction method of multipliers (ADMM) and 2-Patch methods were produced and compared for the individualized P20/N20 somatosensory target for 10 subjects. - Results - D-CMI and MI showed highest intensities parallel to the P20/N20 target compared to ADMM and 2-Patch, with ADMM achieving highest focality. D-CMI showed a slight reduction in intensity compared to MI while reducing side effects and skin level sensations by current distribution over multiple stimulation electrodes. - Conclusion - Individualized D-CMI montages are preferred for our follow up somatosensory experiment to provide a good balance between high current intensities at the target and reduced side effects and skin sensations. - Significance - An integrated combined MEG and EEG source analysis with D-CMI montages for mc-tDCS stimulation potentially can improve control, reproducibility and reduce sensitivity differences between sham and real stimulations.
Experimental demonstration of improved magnetorelaxometry imaging performance using optimized coil configurations. - In: Medical physics, ISSN 2473-4209, Bd. 49 (2022), 5, S. 3361-3374
Background Magnetorelaxometry imaging is an experimental imaging technique capable of reconstructing magnetic nanoparticle distributions inside a volume noninvasively and with high specificity. Thus, magnetorelaxometry imaging is a promising candidate for monitoring a number of therapeutical approaches that employ magnetic nanoparticles, such as magnetic drug targeting and magnetic hyperthermia, to guarantee their safety and efficacy. Prior to a potential clinical application of this imaging modality, it is necessary to optimize magnetorelaxometry imaging systems to produce reliable imaging results and to maximize the reconstruction accuracy of the magnetic nanoparticle distributions. Multiple optimization approaches were already applied throughout a number of simulation studies, all of which yielded increased imaging qualities compared to intuitively designed measurement setups. Purpose None of these simulative approaches was conducted in practice such that it still remains unclear if the theoretical results are achievable in an experimental setting. In this study, we demonstrate the technical feasibility and the increased reconstruction accuracy of optimized coil configurations in two distinct magnetorelaxometry setups. Methods The electromagnetic coil positions and radii of a cuboidal as well as a cylindrical magnetorelaxometry imaging setup are optimized by minimizing the system matrix condition numbers of their corresponding linear forward models. The optimized coil configurations are manufactured alongside with two regular coil grids. Magnetorelaxometry measurements of three cuboidal and four cylindrical magnetic nanoparticle phantoms are conducted, and the resulting reconstruction qualities of the optimized and the regular coil configurations are compared. Results The computed condition numbers of the optimized coil configurations are approximately one order of magnitude lower compared to the regular coil grids. The reconstruction results show that for both setups, every phantom is recovered more accurately by the optimized coil configurations compared to the regular coil grids. Additionally, the optimized coil configurations yield better signal qualities. Conclusions The presented experimental study provides a proof of the practicality and the efficacy of optimizing magnetorelaxometry imaging systems with respect to the condition numbers of their system matrices, previously only demonstrated in simulations. From the promising results of our study, we infer that the minimization of the system matrix condition number will also enable the practical optimization of other design parameters of magnetorelaxometry imaging setups (e.g., sensor configuration, coil currents, etc.) in order to improve the achievable reconstruction qualities even further, eventually paving the way towards clinical application of this imaging modality.
Deep learning for clinical decision support in oncology. - Ilmenau : Universitätsbibliothek, 2022. - 1 Online-Ressource (168 Seiten)
Technische Universität Ilmenau, Dissertation 2022
Bibliography p. 123-139
In den letzten Jahrzehnten sind medizinische Bildgebungsverfahren wie die Computertomographie (CT) zu einem unersetzbaren Werkzeug moderner Medizin geworden, welche eine zeitnahe, nicht-invasive Begutachtung von Organen und Geweben ermöglichen. Die Menge an anfallenden Daten ist dabei rapide gestiegen, allein innerhalb der letzten Jahre um den Faktor 15, und aktuell verantwortlich für 30 % des weltweiten Datenvolumens. Die Anzahl ausgebildeter Radiologen ist weitestgehend stabil, wodurch die medizinische Bildanalyse, angesiedelt zwischen Medizin und Ingenieurwissenschaften, zu einem schnell wachsenden Feld geworden ist. Eine erfolgreiche Anwendung verspricht Zeitersparnisse, und kann zu einer höheren diagnostischen Qualität beitragen. Viele Arbeiten fokussieren sich auf "Radiomics", die Extraktion und Analyse von manuell konstruierten Features. Diese sind jedoch anfällig gegenüber externen Faktoren wie dem Bildgebungsprotokoll, woraus Implikationen für Reproduzierbarkeit und klinische Anwendbarkeit resultieren. In jüngster Zeit sind Methoden des "Deep Learning" zu einer häufig verwendeten Lösung algorithmischer Problemstellungen geworden. Durch Anwendungen in Bereichen wie Robotik, Physik, Mathematik und Wirtschaft, wurde die Forschung im Bereich maschinellen Lernens wesentlich verändert. Ein Kriterium für den Erfolg stellt die Verfügbarkeit großer Datenmengen dar. Diese sind im medizinischen Bereich rar, da die Bilddaten strengen Anforderungen bezüglich Datenschutz und Datensicherheit unterliegen, und oft heterogene Qualität, sowie ungleichmäßige oder fehlerhafte Annotationen aufweisen, wodurch ein bedeutender Teil der Methoden keine Anwendung finden kann. Angesiedelt im Bereich onkologischer Bildgebung zeigt diese Arbeit Wege zur erfolgreichen Nutzung von Deep Learning für medizinische Bilddaten auf. Mittels neuer Methoden für klinisch relevante Anwendungen wie die Schätzung von Läsionswachtum, Überleben, und Entscheidungkonfidenz, sowie Meta-Learning, Klassifikator-Ensembling, und Entscheidungsvisualisierung, werden Wege zur Verbesserungen gegenüber State-of-the-Art-Algorithmen aufgezeigt, welche ein breites Anwendungsfeld haben. Hierdurch leistet die Arbeit einen wesentlichen Beitrag in Richtung einer klinischen Anwendung von Deep Learning, zielt auf eine verbesserte Diagnose, und damit letztlich eine verbesserte Gesundheitsversorgung insgesamt.
Perception thresholds and qualitative perceptions for electrocutaneous stimulation. - In: Scientific reports, ISSN 2045-2322, Bd. 12 (2022), 7335, S. 1-12
Our long-term goal is the development of a wearable warning system that uses electrocutaneous stimulation. To find appropriate stimulation parameters and electrode configurations, we investigate perception amplitude thresholds and qualitative perceptions of electrocutaneous stimulation for varying pulse widths, electrode sizes, and electrode positions. The upper right arm was stimulated in 81 healthy volunteers with biphasic rectangular current pulses varying between 20 and 2000 [my]s. We determined perception, attention, and intolerance thresholds and the corresponding qualitative perceptions for 8 electrode pairs distributed around the upper arm. For a pulse width of 150 [my]s, we find median values of 3.5, 6.9, and 13.8 mA for perception, attention, and intolerance thresholds, respectively. All thresholds decrease with increasing pulse width. Lateral electrode positions have higher intolerance thresholds than medial electrode positions, but perception and attention threshold are not significantly different across electrode positions. Electrode size between 15 × 15 mm^2 and 40 × 40 mm^2 has no significant influence on the thresholds. Knocking is the prevailing perception for perception and attention thresholds while mostly muscle twitching, pinching, and stinging are reported at the intolerance threshold. Biphasic stimulation pulse widths between 150 [my]s and 250 [my]s are suitable for electric warning wearables. Within the given practical limits at the upper arm, electrode size, inter-electrode distance, and electrode position are flexible parameters of electric warning wearables. Our investigations provide the basis for electric warning wearables.