107036 research outputs found
Sort by
Updated prostate cancer risk groups by PSMA-PET PROMISE (PPP2): results from an international multi-centre registry study [Abstract]
No full textO142/#196 Tumor brain organoids for automated drug testing in embryonal brain tumors - on the way to personalized therapy? [Abstract]
No full textSpin-orbit-enabled realization of arbitrary two-qubit gates on moving spins
No full textShuttling spin qubits in systems with large spin-orbit interaction (SOI) can cause errors during motion. However, in this work, we demonstrate that SOI can be harnessed to implement an arbitrary high-fidelity two-qubit (2Q) gate. We consider two spin qubits defined in a semiconductor double quantum dot that are smoothly moved toward each other by gate voltages. We show that an arbitrary high-fidelity 2Q gate can be realized by controlling the shuttling speed and waiting times, and leveraging strong intrinsic or extrinsic SOI. Crucially, performing 2Q operations during qubit transport enables a one-step realization of a wide range of 2Q gates, which often involve several steps when implemented using static dots. Our findings establish a practical route toward direct implementation of any 2Q gate via spin shuttling, significantly reducing control overhead in scalable quantum computing architectures
Treatment of the infection in deep sternal wound infections — a single center cohort study
No full textBackground: In the treatment of deep sternal wound infections (DSWIs) using negative pressure wound therapy (NPWT), infection therapy is separated from reconstruction of the thoracic wall. It is difficult to decide when NPWT can end in order to close the wound. Often ideal conditions cannot be achieved. Must or can sterility be waited for and should antibiotics be administered and if so for how long? These questions are discussed here.
Methods: A retrospective single center analysis is performed. Patients who developed a DSWI after cardiac surgery between 2014 and 2020 are included. The following parameters are collected: baseline data, number of negative pressure dressing changes, frequency of persistence of infection and antibacterial therapy performed at closure. Either a short-term antibiotic therapy over 5 days or a therapy corresponding to the osteomyelitis treatment over 6 weeks was carried out.
Results: In 79 patients, sterility was achieved after 10.2±3.9 (mean ± standard deviation) changes of the negative pressure dressing using NPWT. In 51 patients, NPWT was terminated after 7.4±5.3 changes without achieving sterility. The number of changes varied significantly. If sterility could be achieved, the frequency of infection complications was significantly lower. Long-term antibiotic therapy resulted in significantly fewer infections than short-term therapy.
Conclusions: Microbiological sterility should be achieved if possible. This requires more dressing changes, but the complications are rarer. If it is not possible to wait for sterility, long-term antibiotic therapy is required if osteomyelitis occurs
Identification, prediction, and compensation of hardware-related disturbances in organic computing systems
No full textThe research domain of Organic Computing is concerned with building computing systems that are able to cope with the ever-increasing complexity in Information and Communication Technology. In order to do so, they are intended to freely adapt and organize themselves in an autonomous manner, with respect to their surrounding environment. Organic Computing systems are expected to be robust towards disturbances within their system context. Accordingly, they should be able to maintain a certain degree of functionality even when encountering such disturbances. Lastly, Organic Computing explicitly considers physical hardware in forms of sensors and actuators in order to perceive and manipulate their environment. Despite this inherent focus on real-world environments, OC lacks a designated approach for encountering disturbances that originate from the potential breakdowns of such physical hardware. This research gap is addressed throughout this thesis.
The significance of such physical disturbances for Organic Computing systems is outlined using the current state of research. It is discussed how physical disturbances differ from software-sided disturbances. Especially the fact that the corresponding disturbed hardware may need some sort of maintenance, usually provided by human repair workers, is outlined. Based on the findings, three research questions are formulated.
RQ1 is concerned with identifying physical disturbances, thus, broken components, within the scope of Organic Computing systems. Here, an approach using an adapted version of the XCSF Learning Classifier System is proposed. The internal learning mechanisms of XCSF are altered such that it is able to work with a limited amount of ground truth, in order to limit the amount of necessary human supervision during the learning process. Furthermore, it is adapted for dealing with the notable class imbalance that arises due to the relation between broken and functioning components. Various triggers based on existing data sources are used as an input for the adapted XCSF variant. The approach is evaluated using various scenarios inspired by a data center, as well as a smart factory.
RQ2 focuses on predicting future physical disturbances among the components of an Organic Computing System. The corresponding approach makes use of information gathered from already identified disturbances, as well as mentioned data sources. Iteratively, a pool of data, labeled with the corresponding disturbance information, is created. This data is used to train a prediction pipeline using the Automated Machine Learning framework TPOT. Inspired by online-learning, this pipeline is retrained with newly gathered ata at certain timesteps until it reaches a desired prediction performance. Besides a brief prestudy for assessing different architectural aspects of the algorithm, an evaluation using four freely available datasets is conducted.
RQ3 is intended to bring together the findings from the other two research questions. First of all, it focuses on assessing, or measuring the physical state of OC systems. Secondly, it is concerned with solving current or future physical disturbances. Here, a taxonomy for assessing the various characteristics of physical disturbances and their influence on Organic Computing systems. Two XCSF-based cost functions, used to assess the influence of physical disturbances on an Organic Computing system's overall performance, are presented. Additionally, various Decision Theory-based approaches for prioritizing necessary maintenance actions, taking the cost function into account, are proposed. The corresponding evaluation is conducted on two scenarios that are again loosely based on a data center, as well as a smart factory