107036 research outputs found
Sort by
A spectral relevance analysis approach to pattern recognition of financial time series
No full textUnderstanding patterns in financial time series is crucial for improving prediction accuracy in algorithmic trading and risk management. This paper presents a novel AI-based computer vision approach for classifying financial time series. Historical price sequences are transformed into Gramian Angular Difference Field (GADF) images and fed into a convolutional neural network (CNN) for pattern recognition. To interpret the CNN’s decision-making process, we apply Spectral Relevance Analysis (SpRAy), enabling the identification of distinct clusters based on relevance maps. Clustering the images according to their relevance profiles reveals groups with significantly higher predictive performance compared to the full dataset. The corresponding relevance patterns highlight favorable price movement structures and are identified via the associated clusters
Secure numerical simulations using fully homomorphic encryption
No full textData privacy is a significant concern when using numerical simulations for sensitive information such as medical, financial, or engineering data—especially in untrusted environments like public cloud infrastructures. Fully homomorphic encryption (FHE) offers a promising solution for achieving data privacy by enabling secure computations directly on encrypted data. Aimed at computational scientists, this work explores the viability of FHE-based, privacy-preserving numerical simulations of partial differential equations. The presented approach utilizes the Cheon-Kim-Kim-Song (CKKS) scheme, a widely used FHE method for approximate arithmetic on real numbers. Two Julia packages are introduced, OpenFHE.jl and SecureArithmetic.jl, which wrap the OpenFHE C++ library to provide a convenient interface for secure arithmetic operations. With these tools, the accuracy and performance of key FHE operations in OpenFHE are evaluated, and implementations of finite difference schemes for solving the linear advection equation with encrypted data are demonstrated. The results show that cryptographically secure numerical simulations are possible, but that careful consideration must be given to the computational overhead and the numerical errors introduced by using FHE. An analysis of the algorithmic restrictions imposed by FHE highlights potential challenges and solutions for extending the approach to other models and methods. While it remains uncertain how broadly the approach can be generalized to more complex algorithms due to CKKS limitations, these findings lay the groundwork for further research on privacy-preserving scientific computing
The exact subgraph hierarchy and its vertex-transitive variant for the stable set problem for Paley graphs
No full textThe stability number of a graph, defined as the cardinality of the largest set of pairwise non-adjacent vertices, is NP-hard to compute. The exact subgraph hierarchy (ESH) provides a sequence of increasingly tighter upper bounds on the stability number, starting with the Lovász theta function at the first level and including all exact subgraph constraints of subgraphs of order into the semidefinite program to compute the Lovász theta function at level .
In this paper, we investigate the ESH for Paley graphs, a class of strongly regular, vertex-transitive graphs. We show that for Paley graphs, the bounds obtained from the ESH remain the Lovász theta function up to a certain threshold level, i.e., the bounds of the ESH do not improve up to a certain level.
To overcome this limitation, we introduce the vertex-transitive ESH for the stable set problem for vertex-transitive graphs such as Paley graphs. We prove that this new hierarchy provides upper bounds on the stability number of vertex-transitive graphs that are at least as tight as those obtained from the ESH. Additionally, our computational experiments reveal that the vertex-transitive ESH produces superior bounds compared to the ESH for Paley graphs
Access-site closure with a kaolin-filled pad after intervention with large access sheaths: a retrospective single-center study
No full textThis report summarizes our initial experience with QuikClot Interventional Hemostatic Bandage (QCI) under thrombocyte aggregation inhibition following interventional procedures using 8 French (Fr) catheters or larger, and investigates the feasibility, efficacy, and safety of procedures conducted on 27 patients.
Methods
We performed a retrospective analysis of patients who underwent QCI application after interventional procedures using 8 French (Fr) catheters or larger between 16.02.2021 and 20.01.2025 at a tertiary referral center.
Results
Of the 27 patients included, 20 were male. The mean age was 69 ± 10.2 years, and the mean BMI was 24.6 kg/m2 (± 4.1 kg/m2). 15/27 of the procedures performed were peripheral interventions for peripheral arterial disease. The most common access sheath size was 8 Fr (23/27), and the maximum size was 14 Fr. The technical success rate of the QCI application was 100%. In all patients, a false aneurysm (FA) in need of revision could be excluded via sonography. QCI-associated complication rate was 3.7 %. One surgical revision was performed within the first 30 days. One case of mortality (1/27) was unlikely to be associated with the VCD.
Conclusion
The present study demonstrated that using QCI for access-site closure is easy and effective in combination with manual compression and a pressure bandage, even in cases of large-bore arteriotomies (≥8Fr) and under thrombocyte aggregation inhibition
Sensor-based data-driven differentiation between Parkinson's tremor and essential tremor
No full textDespite distinct pathophysiologies, tremor due to Parkinson’s disease and essential tremor are commonly misdiagnosed due to overlapping symptoms, hindering clinical intervention. Digital phenotyping with machine learning applied to peripheral sensor-based signals has shown potential. However, it lacks validation using an independent dataset or commercial medical device recordings broadly available for routine clinical use. Here, we present a scalable and generalizable diagnostic approach using features engineered from frequency, power, and non-linear phase domains combined with tree-based classification algorithms. Using data from a single hospital, our XGBoost model achieved a diagnostic specificity of 0.92, recall of 1.0, and ROC-AUC (Receiver Operating Characteristic − Area Under the Curve) of 1.0 on a held-out test set. Applying our pre-trained model to an external dataset, recorded in a different clinical setting using other devices, demonstrated strong generalizability with specificity of 0.70, recall of 0.80, and ROC-AUC of 0.79. Our findings not only outperform previous studies in predictive accuracy but also deliver clinically validated metrics that can be implemented in practice. These results highlight the feasibility of deploying affordable, widely available sensor-based diagnostics to enhance clinical accuracy and inform adaptive therapeutic interventions
Verwendung der Larynxmaske bei Adenotonsillektomien im Kindesalter: retrospektive Studie der Daten von 2017–2020
No full textTonsillektomien, Tonsillotomien und Adenotomien (im Folgenden unter Adenotonsillektomien subsumiert) sind alltägliche Eingriffe in der Hals-, Nasen- und Ohrenheilkunde. Im Jahr 2022 wurden in Deutschland 45.378 Tonsillektomien und/oder Adenotomien bei Kindern unter 15 Jahren durchgeführt. (Gesundheitsberichterstattung des Bundes 2024) In den USA werden ungefähr 530.000 ambulante Adenotonsillektomien pro Jahr durchgeführt. (Cullen, Hall, and Golosinskiy 2009) Komplikationslose perioperative Verläufe verkürzen den Heilungsprozess und verringern den Schmerzmittelbedarf.(Sierpina et al. 2012a) Sichere, kurze und komplikationslose Operationen sollten daher das oberste Bestreben sein. Bis 1983 war das Standardverfahren die endotracheale Intubation, welche anschließend durch die eingeführte Larynxmaske schrittweise ersetzt bzw. ergänzt wurde.(Brain 1983) Die Larynxmaske wird ohne die Hilfe eines Laryngoskops eingeführt, wodurch weniger pharyngale Traumata oder Stimulationen, wie vegetative oder respiratorische Reflexe ausgelöst werden.(Aziz and Bashir 2006) Gerade bei pädiatrischen Eingriffen ist eine kurze Beatmungsdauer und die Reduktion der Manipulation an den Atemwegen von höchster Priorität, da die Wahrscheinlichkeit für Komplikationen der Atemwege bei oropharyngealen Operationen signifikant erhöht ist. (Murat, Constant, and Maud’huy 2004; Virag et al. 2019) Patientinnen und Patienten mit einem schweren obstruktiven Schlafapnoesyndrom (OSAS) haben ein fünffach höheres Risiko respiratorische Komplikationen zu erleiden.(De Luca Canto et al. 2015) Es gibt mehrere Berichte, dass Kinder nach Tonsillektomien aufgrund von anästhesiologischen Komplikationen, wie Opioidüberhang oder Apnoe, versterben oder neurologische Schäden erleiden.(Coté, Posner, and Domino 2014) Aufgrund der geringeren Manipulation an den vulnerablen kindlichen Atemwegen erwies sich die Larynxmaske in mehreren Studien bei kurzen Eingriffen als eine gleichwertige oder zumindest nicht unterlegene Atemwegssicherung im Vergleich zum Endotrachealtubus. (Karin Becke 2014a)
Beide Beatmungsmethoden haben sowohl anästhesiologisch wie auch chirurgisch Vor- und Nachteile. Bisherige vereinzelte Studien zu Beatmungsmethoden in pädiatrischen HNO-Operationen widersprachen sich teilweise oder führten zu keiner klaren Aussage,
weshalb das Ziel dieser Arbeit sein soll, statistisch anhand 826 Operationen retrospektiv die Beatmungswege zu analysieren und aufzuarbeiten.
Die Klinik für Anästhesiologie und operative Intensivmedizin des Universitätsklinikum Augsburg betreut Adenotomien und Tonsillotomien bei Kleinkindern unter drei Jahren und bei vorerkrankten Kindern am Standort Universitätsklinikum Augsburg. Kinder über 3 Jahren werden am Medizincampus Süd in Augsburg ambulant behandelt, da es dort keine pädiatrische intensivmedizinische Station gibt. Dabei wird die anästhesiologische Vorgehensweise von dem jeweils behandelnden Anästhesisten tagesaktuell in Rücksprache mit dem jeweiligen Operateur festgelegt. Im Universitätsklinikum Augsburg wird vermehrt (n = 502; 60,8 %) die Larynxmaske zur Atemwegssicherung während des Eingriffs bevorzugt.
Die Analyse von OP-Zeiten, Komplikationsraten und Outcome-Daten soll zeigen, ob eine Empfehlung zugunsten einer Intubation oder einer Larynxmaske erarbeitet werden kann, und ob man so Komplikationen in Verbindung mit Risikofaktoren wie bspw. BMI oder ASA-Status vorbeugen kann.
Daten von 826 Operationen wurden für den Zeitraum von April 2017 bis Februar 2020 ausgewertet. Dabei wurden zu jeder Operation 32 Parameter erhoben, welche in Kapitel 5.3.1 aufgelistet sind
Does acqui-hiring pay off? An empirical investigation of founder retention
No full textIn the fiercely competitive global talent landscape, even tech giants struggle to attract and retain top-tier entrepreneurial talent. To address this challenge, companies increasingly employ acqui-hiring—a strategic approach based on mergers and acquisitions aimed at acquiring startups primarily for their human capital. However, empirical evidence supporting the effectiveness of acqui-hiring is limited and contentious. Our study aims to fill this gap by investigating the retention decisions of acqui-hired founders, who are central to the success of such acquisitions. Leveraging a dataset comprising 454 acqui-hired founders from 241 transactions by Google and Meta (Facebook), our findings highlight the difficulty of retaining individual entrepreneurs compared to entire co-founder teams, and underscore the importance of maintaining co-founders’ hierarchy. Our research contributes to the strategic entrepreneurship literature by exploring nuanced organizational design decisions in intrapreneurship and startup–corporate collaborations, shedding light on the efficacy of structural integration for knowledge and innovation transfer
The case of fleeting orders and flickering quotes
No full textThe literature controversially discusses the ambiguous motives and driving forces behind quickly cancelled limit orders (fleeting orders), which are characteristic of high-frequency markets. In particular, manipulative and dysfunctional characteristics are feared. We analyze top-of-book fleeting orders—so-called flickering quotes—and show with an ultra-low latency derivative data set that none of these properties have to be dreaded. On the contrary, flickering quotes are associated with liquid market environments, for example: the prices of “flickering” order books improve by 3.90% before trades. The results reveal that flickering quotes are likely due to beneficial price discovery processes. Additionally, HFTs might offer their excess positions at a discount to other participants with these orders
