Hasso-Plattner-Institut für Digital Engineering gGmbH
Refine
Year of publication
Document Type
- Article (195)
- Monograph/Edited Volume (122)
- Doctoral Thesis (46)
- Other (31)
- Conference Proceeding (12)
- Preprint (4)
- Postprint (2)
- Part of a Book (1)
- Review (1)
Language
- English (372)
- German (40)
- Multiple languages (2)
Keywords
- Cloud Computing (9)
- cloud computing (8)
- Hasso-Plattner-Institut (7)
- Datenintegration (6)
- Forschungskolleg (6)
- Hasso Plattner Institute (6)
- Klausurtagung (6)
- Modellierung (6)
- Service-oriented Systems Engineering (6)
- Forschungsprojekte (5)
Fast style transfer methods have recently gained popularity in art-related applications as they make a generalized real-time stylization of images practicable.
However, they are mostly limited to one-shot stylizations concerning the interactive adjustment of style elements.
In particular, the expressive control over stroke sizes or stroke orientations remains an open challenge.
To this end, we propose a novel stroke-adjustable fast style transfer network that enables simultaneous control over the stroke size and intensity, and allows a wider range of expressive editing than current approaches by utilizing the scale-variance of convolutional neural networks.
Furthermore, we introduce a network-agnostic approach for style-element editing by applying reversible input transformations that can adjust strokes in the stylized output. At this, stroke orientations can be adjusted, and warping-based effects can be applied to stylistic elements, such as swirls or waves.
To demonstrate the real-world applicability of our approach, we present StyleTune, a mobile app for interactive editing of neural style transfers at multiple levels of control. Our app allows stroke adjustments on a global and local level.
It furthermore implements an on-device patch-based upsampling step that enables users to achieve results with high output fidelity and resolutions of more than 20 megapixels.
Our approach allows users to art-direct their creations and achieve results that are not possible with current style transfer applications.
Detecting anomalous subsequences in time series is an important task in time series analytics because it serves the identification of special events, such as production faults, delivery bottlenecks, system defects, or heart flicker.
Consequently, many algorithms have been developed for the automatic detection of such anomalous patterns. The enormous number of approaches (i.e., more than 158 as of today), the lack of properly labeled test data, and the complexity of time series anomaly benchmarking have, though, led to a situation where choosing the best detection technique for a given anomaly detection task is a difficult challenge.
In this demonstration, we present TIMEEVAL, an extensible, scalable and automatic benchmarking toolkit for time series anomaly detection algorithms. TIMEEVAL includes an extensive data generator and supports both interactive and batch evaluation scenarios. With our novel toolkit, we aim to ease the evaluation effort and help the community to provide more meaningful evaluations.
Entity resolution on-demand
(2022)
Entity Resolution (ER) aims to identify and merge records that refer to the same real-world entity.
ER is typically employed as an expensive cleaning step on the entire data before consuming it. Yet, determining which entities are useful once cleaned depends solely on the user's application, which may need only a fraction of them.
For instance, when dealing with Web data, we would like to be able to filter the entities of interest gathered from multiple sources without cleaning the entire, continuously-growing data.
Similarly, when querying data lakes, we want to transform data on-demand and return the results in a timely manner-a fundamental requirement of ELT (Extract-Load-Transform) pipelines.
We propose BrewER, a framework to evaluate SQL SP queries on dirty data while progressively returning results as if they were issued on cleaned data. BrewER tries to focus the cleaning effort on one entity at a time, following an ORDER BY predicate. Thus, it inherently supports top-k and stop-and-resume execution.
For a wide range of applications, a significant amount of resources can be saved. We exhaustively evaluate and show the efficacy of BrewER on four real-world datasets.
Persistent memory's (PMem) byte-addressability and persistence at DRAM-like speed with SSD-like capacity have the potential to cause a major performance shift in database storage systems. With the availability of Intel Optane DC Persistent Memory, initial benchmarks evaluate the performance of real PMem hardware.
However, these results apply to only a single server and it is not yet clear how workloads compare across different PMem servers. In this paper, we propose PerMA-Bench, a con.gurable benchmark framework that allows users to evaluate the bandwidth, latency, and operations per second for customizable database-related PMem access.
Based on PerMA-Bench, we perform an extensive evaluation of PMem performance across four di.erent server configurations, containing both first- and second-generation Optane, with additional parameters such as DIMM power budget and number of DIMMs per server.
We validate our results with existing systems and show the impact of low-level design choices. We conduct a price-performance comparison that shows while there are large differences across Optane DIMMs, PMem is generally competitive with DRAM. We discuss our findings and identify eight general and implementation-specific aspects that influence PMem performance and should be considered in future work to improve PMem-aware designs.
HARE
(2023)
Sensor-based human activity recognition is becoming ever more prevalent. The increasing importance of distinguishing human movements, particularly in healthcare, coincides with the advent of increasingly compact sensors. A complex sequence of individual steps currently characterizes the activity recognition pipeline. It involves separate data collection, preparation, and processing steps, resulting in a heterogeneous and fragmented process. To address these challenges, we present a comprehensive framework, HARE, which seamlessly integrates all necessary steps. HARE offers synchronized data collection and labeling, integrated pose estimation for data anonymization, a multimodal classification approach, and a novel method for determining optimal sensor placement to enhance classification results. Additionally, our framework incorporates real-time activity recognition with on-device model adaptation capabilities. To validate the effectiveness of our framework, we conducted extensive evaluations using diverse datasets, including our own collected dataset focusing on nursing activities. Our results show that HARE’s multimodal and on-device trained model outperforms conventional single-modal and offline variants. Furthermore, our vision-based approach for optimal sensor placement yields comparable results to the trained model. Our work advances the field of sensor-based human activity recognition by introducing a comprehensive framework that streamlines data collection and classification while offering a novel method for determining optimal sensor placement.
DUO-GAIT
(2023)
In recent years, there has been a growing interest in developing and evaluating gait analysis algorithms based on inertial measurement unit (IMU) data, which has important implications, including sports, assessment of diseases, and rehabilitation. Multi-tasking and physical fatigue are two relevant aspects of daily life gait monitoring, but there is a lack of publicly available datasets to support the development and testing of methods using a mobile IMU setup. We present a dataset consisting of 6-minute walks under single- (only walking) and dual-task (walking while performing a cognitive task) conditions in unfatigued and fatigued states from sixteen healthy adults. Especially, nine IMUs were placed on the head, chest, lower back, wrists, legs, and feet to record under each of the above-mentioned conditions. The dataset also includes a rich set of spatio-temporal gait parameters that capture the aspects of pace, symmetry, and variability, as well as additional study-related information to support further analysis. This dataset can serve as a foundation for future research on gait monitoring in free-living environments.
Background: Wearable multi-modal time-series classification applications outperform their best uni-modal counterparts and hold great promise. A modality that directly measures electrical correlates from the brain is electroencephalography. Due to varying noise sources, different key brain regions, key frequency bands, and signal characteristics like non-stationarity, techniques for data pre-processing and classification algorithms are task-dependent.
Method: Here, a systematic literature review on mental state classification for wearable electroencephalog-raphy is presented. Four search terms in different combinations were used for an in-title search. The search was executed on the 29th of June 2022, across Google Scholar, PubMed, IEEEXplore, and ScienceDirect. 76 most relevant publications were set into context as the current state-of-the-art in mental state time-series classification.
Results: Pre-processing techniques, features, and time-series classification models were analyzed. Across publications, a window length of one second was mainly chosen for classification and spectral features were utilized the most. The achieved performance per time-series classification model is analyzed, finding linear discriminant analysis, decision trees, and k-nearest neighbors models outperform support-vector machines by a factor of up to 1.5. A historical analysis depicts future trends while under-reported aspects relevant to practical applications are discussed.
Conclusions: Five main conclusions are given, covering utilization of available area for electrode placement on the head, most often or scarcely utilized features and time-series classification model architectures, baseline reporting practices, as well as explainability and interpretability of Deep Learning. The importance of a 'test battery' assessing the influence of data pre-processing and multi-modality on time-series classification performance is emphasized.
Pressure overload in patients with aortic valve stenosis and volume overload in mitral valve regurgitation trigger specific forms of cardiac remodeling; however, little is known about similarities and differences in myocardial proteome regulation. We performed proteome profiling of 75 human left ventricular myocardial biopsies (aortic stenosis = 41, mitral regurgitation = 17, and controls = 17) using high-resolution tandem mass spectrometry next to clinical and hemodynamic parameter acquisition. In patients of both disease groups, proteins related to ECM and cytoskeleton were more abundant, whereas those related to energy metabolism and proteostasis were less abundant compared with controls. In addition, disease group-specific and sex-specific differences have been observed. Male patients with aortic stenosis showed more proteins related to fibrosis and less to energy metabolism, whereas female patients showed strong reduction in proteostasis-related proteins. Clinical imaging was in line with proteomic findings, showing elevation of fibrosis in both patient groups and sex differences. Disease-and sex-specific proteomic profiles provide insight into cardiac remodeling in patients with heart valve disease and might help improve the understanding of molecular mechanisms and the development of individualized treatment strategies.
High annotation costs are a substantial bottleneck in applying deep learning architectures to clinically relevant use cases, substantiating the need for algorithms to learn from unlabeled data.
In this work, we propose employing self-supervised methods. To that end, we trained with three self-supervised algorithms on a large corpus of unlabeled dental images, which contained 38K bitewing radiographs (BWRs). We then applied the learned neural network representations on tooth-level dental caries classification, for which we utilized labels extracted from electronic health records (EHRs). Finally, a holdout test-set was established, which consisted of 343 BWRs and was annotated by three dental professionals and approved by a senior dentist.
This test-set was used to evaluate the fine-tuned caries classification models. Our experimental results demonstrate the obtained gains by pretraining models using self-supervised algorithms. These include improved caries classification performance (6 p.p. increase in sensitivity) and, most importantly, improved label-efficiency.
In other words, the resulting models can be fine-tuned using few labels (annotations).
Our results show that using as few as 18 annotations can produce >= 45% sensitivity, which is comparable to human-level diagnostic performance.
This study shows that self-supervision can provide gains in medical image analysis, particularly when obtaining labels is costly and expensive.
Fragmentation of peptides leaves characteristic patterns in mass spectrometry data, which can be used to identify protein sequences, but this method is challenging for mutated or modified sequences for which limited information exist. Altenburg et al. use an ad hoc learning approach to learn relevant patterns directly from unannotated fragmentation spectra.
Mass spectrometry-based proteomics provides a holistic snapshot of the entire protein set of living cells on a molecular level. Currently, only a few deep learning approaches exist that involve peptide fragmentation spectra, which represent partial sequence information of proteins.
Commonly, these approaches lack the ability to characterize less studied or even unknown patterns in spectra because of their use of explicit domain knowledge.
Here, to elevate unrestricted learning from spectra, we introduce 'ad hoc learning of fragmentation' (AHLF), a deep learning model that is end-to-end trained on 19.2 million spectra from several phosphoproteomic datasets. AHLF is interpretable, and we show that peak-level feature importance values and pairwise interactions between peaks are in line with corresponding peptide fragments.
We demonstrate our approach by detecting post-translational modifications, specifically protein phosphorylation based on only the fragmentation spectrum without a database search. AHLF increases the area under the receiver operating characteristic curve (AUC) by an average of 9.4% on recent phosphoproteomic data compared with the current state of the art on this task.
Furthermore, use of AHLF in rescoring search results increases the number of phosphopeptide identifications by a margin of up to 15.1% at a constant false discovery rate. To show the broad applicability of AHLF, we use transfer learning to also detect cross-linked peptides, as used in protein structure analysis, with an AUC of up to 94%.
N-of-1 trials are the gold standard study design to evaluate individual treatment effects and derive personalized treatment strategies. Digital tools have the potential to initiate a new era of N-of-1 trials in terms of scale and scope, but fully functional platforms are not yet available.
Here, we present the open source StudyU platform, which includes the StudyU Designer and StudyU app.
With the StudyU Designer, scientists are given a collaborative web application to digitally specify, publish, and conduct N-of-1 trials.
The StudyU app is a smartphone app with innovative user-centric elements for participants to partake in trials published through the StudyU Designer to assess the effects of different interventions on their health.
Thereby, the StudyU platform allows clinicians and researchers worldwide to easily design and conduct digital N-of-1 trials in a safe manner.
We envision that StudyU can change the landscape of personalized treatments both for patients and healthy individuals, democratize and personalize evidence generation for self-optimization and medicine, and can be integrated in clinical practice.
The investigation of metabolic fluxes and metabolite distributions within cells by means of tracer molecules is a valuable tool to unravel the complexity of biological systems. Technological advances in mass spectrometry (MS) technology such as atmospheric pressure chemical ionization (APCI) coupled with high resolution (HR), not only allows for highly sensitive analyses but also broadens the usefulness of tracer-based experiments, as interesting signals can be annotated de novo when not yet present in a compound library. However, several effects in the APCI ion source, i.e., fragmentation and rearrangement, lead to superimposed mass isotopologue distributions (MID) within the mass spectra, which need to be corrected during data evaluation as they will impair enrichment calculation otherwise. Here, we present and evaluate a novel software tool to automatically perform such corrections. We discuss the different effects, explain the implemented algorithm, and show its application on several experimental datasets. This adjustable tool is available as an R package from CRAN.
Ion-mobility spectrometry shows great promise to tackle analytically challenging research questions by adding another separation dimension to liquid chromatography-mass spectrometry.
The understanding of how analyte properties influence ion mobility has increased through recent studies, but no clear rationale for the design of customized experimental settings has emerged.
Here, we leverage machine learning to deepen our understanding of field asymmetric waveform ion-mobility spectrometry for the analysis of cross-linked peptides.
Knowing that predominantly m/z and then the size and charge state of an analyte influence the separation, we found ideal compensation voltages correlating with the size exclusion chromatography fraction number.
The effect of this relationship on the analytical depth can be substantial as exploiting it allowed us to almost double unique residue pair detections in a proteome-wide cross-linking experiment.
Other applications involving liquid- and gas-phase separation may also benefit from considering such parameter dependencies.
Modern data analysis tasks often involve control flow statements, such as the iterations in PageRank and K-means. To achieve scalability, developers usually implement these tasks in distributed dataflow systems, such as Spark and Flink. Designers of such systems have to choose between providing imperative or functional control flow constructs to users. Imperative constructs are easier to use, but functional constructs are easier to compile to an efficient dataflow job. We propose Mitos, a system where control flow is both easy to use and efficient. Mitos relies on an intermediate representation based on the static single assignment form. This allows us to abstract away from specific control flow constructs and treat any imperative control flow uniformly both when building the dataflow job and when coordinating the distributed execution.
In his article, 'Social constructionism and climate science denial', Hansson claims to present empirical evidence that the cultural theory developed by Dame Mary Douglas, Aaron Wildavsky and ourselves (among others) leads to (climate) science denial. In this reply, we show that there is no validity to these claims. First, we show that Hansson's empirical evidence that cultural theory has led to climate science denial falls apart under closer inspection. Contrary to Hansson's claims, cultural theory has made significant contributions to understanding and addressing climate change. Second, we discuss various features of Douglas' cultural theory that differentiate it from other constructivist approaches and make it compatible with the scientific method. Thus, we also demonstrate that cultural theory cannot be accused of epistemic relativism.
Data encoding has been applied to database systems for decades as it mitigates bandwidth bottlenecks and reduces storage requirements. But even in the presence of these advantages, most in-memory database systems use data encoding only conservatively as the negative impact on runtime performance can be severe. Real-world systems with large parts being infrequently accessed and cost efficiency constraints in cloud environments require solutions that automatically and efficiently select encoding techniques, including heavy-weight compression. In this paper, we introduce workload-driven approaches to automaticaly determine memory budget-constrained encoding configurations using greedy heuristics and linear programming. We show for TPC-H, TPC-DS, and the Join Order Benchmark that optimized encoding configurations can reduce the main memory footprint significantly without a loss in runtime performance over state-of-the-art dictionary encoding. To yield robust selections, we extend the linear programming-based approach to incorporate query runtime constraints and mitigate unexpected performance regressions.
How inclusive are we?
(2022)
ACM SIGMOD, VLDB and other database organizations have committed to fostering an inclusive and diverse community, as do many other scientific organizations. Recently, different measures have been taken to advance these goals, especially for underrepresented groups. One possible measure is double-blind reviewing, which aims to hide gender, ethnicity, and other properties of the authors. <br /> We report the preliminary results of a gender diversity analysis of publications of the database community across several peer-reviewed venues, and also compare women's authorship percentages in both single-blind and double-blind venues along the years. We also obtained a cross comparison of the obtained results in data management with other relevant areas in Computer Science.
Fast style transfer methods have recently gained popularity in art-related applications as they make a generalized real-time stylization of images practicable. However, they are mostly limited to one-shot stylizations concerning the interactive adjustment of style elements. In particular, the expressive control over stroke sizes or stroke orientations remains an open challenge. To this end, we propose a novel stroke-adjustable fast style transfer network that enables simultaneous control over the stroke size and intensity, and allows a wider range of expressive editing than current approaches by utilizing the scale-variance of convolutional neural networks. Furthermore, we introduce a network-agnostic approach for style-element editing by applying reversible input transformations that can adjust strokes in the stylized output. At this, stroke orientations can be adjusted, and warping-based effects can be applied to stylistic elements, such as swirls or waves. To demonstrate the real-world applicability of our approach, we present StyleTune, a mobile app for interactive editing of neural style transfers at multiple levels of control. Our app allows stroke adjustments on a global and local level. It furthermore implements an on-device patch-based upsampling step that enables users to achieve results with high output fidelity and resolutions of more than 20 megapixels. Our approach allows users to art-direct their creations and achieve results that are not possible with current style transfer applications.
Correction to: Knowledge bases and software support for variant interpretation in precision oncology
(2021)
Precision oncology is a rapidly evolving interdisciplinary medical specialty. Comprehensive cancer panels are becoming increasingly available at pathology departments worldwide, creating the urgent need for scalable cancer variant annotation and molecularly informed treatment recommendations. A wealth of mainly academia-driven knowledge bases calls for software tools supporting the multi-step diagnostic process. We derive a comprehensive list of knowledge bases relevant for variant interpretation by a review of existing literature followed by a survey among medical experts from university hospitals in Germany. In addition, we review cancer variant interpretation tools, which integrate multiple knowledge bases. We categorize the knowledge bases along the diagnostic process in precision oncology and analyze programmatic access options as well as the integration of knowledge bases into software tools. The most commonly used knowledge bases provide good programmatic access options and have been integrated into a range of software tools. For the wider set of knowledge bases, access options vary across different parts of the diagnostic process. Programmatic access is limited for information regarding clinical classifications of variants and for therapy recommendations. The main issue for databases used for biological classification of pathogenic variants and pathway context information is the lack of standardized interfaces. There is no single cancer variant interpretation tool that integrates all identified knowledge bases. Specialized tools are available and need to be further developed for different steps in the diagnostic process.
In clinical practice, only a few reliable measurement instruments are available for monitoring knee joint rehabilitation. Advances to replace motion capturing with sensor data measurement have been made in the last years. Thus, a systematic review of the literature was performed, focusing on the implementation, diagnostic accuracy, and facilitators and barriers of integrating wearable sensor technology in clinical practices based on a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. For critical appraisal, the COSMIN Risk of Bias tool for reliability and measurement of error was used. PUBMED, Prospero, Cochrane database, and EMBASE were searched for eligible studies. Six studies reporting reliability aspects in using wearable sensor technology at any point after knee surgery in humans were included. All studies reported excellent results with high reliability coefficients, high limits of agreement, or a few detectable errors. They used different or partly inappropriate methods for estimating reliability or missed reporting essential information. Therefore, a moderate risk of bias must be considered. Further quality criterion studies in clinical settings are needed to synthesize the evidence for providing transparent recommendations for the clinical use of wearable movement sensors in knee joint rehabilitation.
FIBER
(2021)
Objectives:
The development of clinical predictive models hinges upon the availability of comprehensive clinical data. Tapping into such resources requires considerable effort from clinicians, data scientists, and engineers. Specifically, these efforts are focused on data extraction and preprocessing steps required prior to modeling, including complex database queries. A handful of software libraries exist that can reduce this complexity by building upon data standards. However, a gap remains concerning electronic health records (EHRs) stored in star schema clinical data warehouses, an approach often adopted in practice. In this article, we introduce the FlexIBle EHR Retrieval (FIBER) tool: a Python library built on top of a star schema (i2b2) clinical data warehouse that enables flexible generation of modeling-ready cohorts as data frames.
Materials and Methods:
FIBER was developed on top of a large-scale star schema EHR database which contains data from 8 million patients and over 120 million encounters. To illustrate FIBER's capabilities, we present its application by building a heart surgery patient cohort with subsequent prediction of acute kidney injury (AKI) with various machine learning models.
Results:
Using FIBER, we were able to build the heart surgery cohort (n = 12 061), identify the patients that developed AKI (n = 1005), and automatically extract relevant features (n = 774). Finally, we trained machine learning models that achieved area under the curve values of up to 0.77 for this exemplary use case.
Conclusion:
FIBER is an open-source Python library developed for extracting information from star schema clinical data warehouses and reduces time-to-modeling, helping to streamline the clinical modeling process.
Phe2vec
(2021)
Robust phenotyping of patients from electronic health records (EHRs) at scale is a challenge in clinical informatics. Here, we introduce Phe2vec, an automated framework for disease phenotyping from EHRs based on unsupervised learning and assess its effectiveness against standard rule-based algorithms from Phenotype KnowledgeBase (PheKB). Phe2vec is based on pre-computing embeddings of medical concepts and patients' clinical history. Disease phenotypes are then derived from a seed concept and its neighbors in the embedding space. Patients are linked to a disease if their embedded representation is close to the disease phenotype. Comparing Phe2vec and PheKB cohorts head-to-head using chart review, Phe2vec performed on par or better in nine out of ten diseases. Differently from other approaches, it can scale to any condition and was validated against widely adopted expert-based standards. Phe2vec aims to optimize clinical informatics research by augmenting current frameworks to characterize patients by condition and derive reliable disease cohorts.
Background:
Childhood and adolescence are critical stages of life for mental health and well-being. Schools are a key setting for mental health promotion and illness prevention. One in five children and adolescents have a mental disorder, about half of mental disorders beginning before the age of 14. Beneficial and explainable artificial intelligence can replace current paper- based and online approaches to school mental health surveys. This can enhance data acquisition, interoperability, data driven analysis, trust and compliance. This paper presents a model for using chatbots for non-obtrusive data collection and supervised machine learning models for data analysis; and discusses ethical considerations pertaining to the use of these models.
Methods:
For data acquisition, the proposed model uses chatbots which interact with students. The conversation log acts as the source of raw data for the machine learning. Pre-processing of the data is automated by filtering for keywords and phrases.
Existing survey results, obtained through current paper-based data collection methods, are evaluated by domain experts (health professionals). These can be used to create a test dataset to validate the machine learning models. Supervised learning
can then be deployed to classify specific behaviour and mental health patterns.
Results:
We present a model that can be used to improve upon current paper-based data collection and manual data analysis methods. An open-source GitHub repository contains necessary tools and components of this model. Privacy is respected through
rigorous observance of confidentiality and data protection requirements. Critical reflection on these ethics and law aspects is included in the project.
Conclusions:
This model strengthens mental health surveillance in schools. The same tools and components could be applied to other public health data. Future extensions of this model could also incorporate unsupervised learning to find clusters and patterns
of unknown effects.
Background:
Early reports indicate that AKI is common among patients with coronavirus disease 2019 (COVID-19) and associatedwith worse outcomes. However, AKI among hospitalized patients with COVID19 in the United States is not well described.
Methods:
This retrospective, observational study involved a review of data from electronic health records of patients aged >= 18 years with laboratory-confirmed COVID-19 admitted to the Mount Sinai Health System from February 27 to May 30, 2020. We describe the frequency of AKI and dialysis requirement, AKI recovery, and adjusted odds ratios (aORs) with mortality.
Results:
Of 3993 hospitalized patients with COVID-19, AKI occurred in 1835 (46%) patients; 347 (19%) of the patientswith AKI required dialysis. The proportionswith stages 1, 2, or 3 AKIwere 39%, 19%, and 42%, respectively. A total of 976 (24%) patients were admitted to intensive care, and 745 (76%) experienced AKI. Of the 435 patients with AKI and urine studies, 84% had proteinuria, 81% had hematuria, and 60% had leukocyturia. Independent predictors of severe AKI were CKD, men, and higher serum potassium at admission. In-hospital mortality was 50% among patients with AKI versus 8% among those without AKI (aOR, 9.2; 95% confidence interval, 7.5 to 11.3). Of survivors with AKI who were discharged, 35% had not recovered to baseline kidney function by the time of discharge. An additional 28 of 77 (36%) patients who had not recovered kidney function at discharge did so on posthospital follow-up.
Conclusions:
AKI is common among patients hospitalized with COVID-19 and is associated with high mortality. Of all patients with AKI, only 30% survived with recovery of kidney function by the time of discharge.
Scientific sexism
(2021)
OBJECTIVE:
To investigate gender inequity in the scientific production of the University of Sao Paulo.
METHODS:
Members of the University of Sao Paulo faculty are the study population. The Web of Science repository was the source of the publication metrics. We selected the measures: total publications and citations, average of citations per year and item, H-index, and history of citations between 1950 and 2019. We used the name of the faculty member as a proxy to the gender identity. We use descriptive statistics to characterize the metrics. We evaluated the scissors effect by selecting faculty members with a high H-index. The historical series of citations was projected until 2100. We carry out analyses for the general population and working time subgroups: less than 10 years, 10 to 20 years, and 20 years or more.
RESULTS:
Of the 8,325 faculty members, we included 3,067 (36.8%). Among those included, 1,893 (61.7%) were male and 1,174 (38.28%) female. The male gender presented higher values in the publication metrics (average of articles: M = 67.0 versus F = 49.7; average of citations/year: M = 53.9 versus F = 35.9), and H-index (M = 14.5 versus F = 12.4). Among the 100 individuals with the highest H-index (>= 37), 83% are male. The male curve grows faster in the historical series of citations, opening a difference between the groups whose separation is confirmed by the projection.
DISCUSSION:
Scientific production at the Universidade de Sao Paulo is subject to a gender bias. Two-thirds of the faculty are male, and hiring over the past few decades perpetuates this pattern. The large majority of high impact faculty members are male.
CONCLUSION:
Our analysis suggests that the Universidade de Sao Paulo will not overcome gender inequality in scientific production without substantive affirmative action. Development does not happen by chance but through choices that are affirmative, decisive, and long-term oriented.
Recently, substantial research effort has focused on how to apply CNNs or RNNs to better capture temporal patterns in videos, so as to improve the accuracy of video classification. In this paper, we investigate the potential of a purely attention based local feature integration. Accounting for the characteristics of such features in video classification, we first propose Basic Attention Clusters (BAC), which concatenates the output of multiple attention units applied in parallel, and introduce a shifting operation to capture more diverse signals. Experiments show that BAC can achieve excellent results on multiple datasets. However, BAC treats all feature channels as an indivisible whole, which is suboptimal for achieving a finer-grained local feature integration over the channel dimension. Additionally, it treats the entire local feature sequence as an unordered set, thus ignoring the sequential relationships. To improve over BAC, we further propose the channel pyramid attention schema by splitting features into sub-features at multiple scales for coarse-to-fine sub-feature interaction modeling, and propose the temporal pyramid attention schema by dividing the feature sequences into ordered sub-sequences of multiple lengths to account for the sequential order. Our final model pyramidxpyramid attention clusters (PPAC) combines both channel pyramid attention and temporal pyramid attention to focus on the most important sub-features, while also preserving the temporal information of the video. We demonstrate the effectiveness of PPAC on seven real-world video classification datasets. Our model achieves competitive results across all of these, showing that our proposed framework can consistently outperform the existing local feature integration methods across a range of different scenarios.
Business processes are often specified in descriptive or normative models. Both types of models should adhere to internal and external regulations, such as company guidelines or laws. Employing compliance checking techniques, it is possible to verify process models against rules. While traditionally compliance checking focuses on well-structured processes, we address case management scenarios. In case management, knowledge workers drive multi-variant and adaptive processes. Our contribution is based on the fragment-based case management approach, which splits a process into a set of fragments. The fragments are synchronized through shared data but can, otherwise, be dynamically instantiated and executed. We formalize case models using Petri nets. We demonstrate the formalization for design-time and run-time compliance checking and present a proof-of-concept implementation. The application of the implemented compliance checking approach to a use case exemplifies its effectiveness while designing a case model. The empirical evaluation on a set of case models for measuring the performance of the approach shows that rules can often be checked in less than a second.
One of the first and easy to use techniques for proving run time bounds for evolutionary algorithms is the so-called method of fitness levels by Wegener. It uses a partition of the search space into a sequence of levels which are traversed by the algorithm in increasing order, possibly skipping levels. An easy, but often strong upper bound for the run time can then be derived by adding the reciprocals of the probabilities to leave the levels (or upper bounds for these). Unfortunately, a similarly effective method for proving lower bounds has not yet been established. The strongest such method, proposed by Sudholt (2013), requires a careful choice of the viscosity parameters gamma(i), j, 0 <= i < j <= n. In this paper we present two new variants of the method, one for upper and one for lower bounds. Besides the level leaving probabilities, they only rely on the probabilities that levels are visited at all. We show that these can be computed or estimated without greater difficulties and apply our method to reprove the following known results in an easy and natural way. (i) The precise run time of the (1+1) EA on LEADINGONES. (ii) A lower bound for the run time of the (1+1) EA on ONEMAX, tight apart from an O(n) term. (iii) A lower bound for the run time of the (1+1) EA on long k-paths (which differs slightly from the previous result due to a small error in the latter). We also prove a tighter lower bound for the run time of the (1+1) EA on jump functions by showing that, regardless of the jump size, only with probability O(2(-n)) the algorithm can avoid to jump over the valley of low fitness.
A standard approach to accelerating shortest path algorithms on networks is the bidirectional search, which explores the graph from the start and the destination, simultaneously. In practice this strategy performs particularly well on scale-free real-world networks. Such networks typically have a heterogeneous degree distribution (e.g., a power-law distribution) and high clustering (i.e., vertices with a common neighbor are likely to be connected themselves). These two properties can be obtained by assuming an underlying hyperbolic geometry. <br /> To explain the observed behavior of the bidirectional search, we analyze its running time on hyperbolic random graphs and prove that it is (O) over tilde (n(2-1/alpha) + n(1/(2 alpha)) + delta(max)) with high probability, where alpha is an element of (1/2, 1) controls the power-law exponent of the degree distribution, and dmax is the maximum degree. This bound is sublinear, improving the obvious worst-case linear bound. Although our analysis depends on the underlying geometry, the algorithm itself is oblivious to it.
A simplified run time analysis of the univariate marginal distribution algorithm on LeadingOnes
(2021)
With elementary means, we prove a stronger run time guarantee for the univariate marginal distribution algorithm (UMDA) optimizing the LEADINGONES benchmark function in the desirable regime with low genetic drift. If the population size is at least quasilinear, then, with high probability, the UMDA samples the optimum in a number of iterations that is linear in the problem size divided by the logarithm of the UMDA's selection rate. This improves over the previous guarantee, obtained by Dang and Lehre (2015) via the deep level-based population method, both in terms of the run time and by demonstrating further run time gains from small selection rates. Under similar assumptions, we prove a lower bound that matches our upper bound up to constant factors.
The engineering of digital twins and their user interaction parts with explicated processes, namely processaware digital twin cockpits (PADTCs), is challenging due to the complexity of the systems and the need for information from different disciplines within the engineering process. Therefore, it is interesting to investigate how to facilitate their engineering by using already existing data, namely event logs, and reducing the number of manual steps for their engineering. Current research lacks systematic, automated approaches to derive process-aware digital twin cockpits even though some helpful techniques already exist in the areas of process mining and software engineering. Within this paper, we present a low-code development approach that reduces the amount of hand-written code needed and uses process mining techniques to derive PADTCs. We describe what models could be derived from event log data, which generative steps are needed for the engineering of PADTCs, and how process mining could be incorporated into the resulting application. This process is evaluated using the MIMIC III dataset for the creation of a PADTC prototype for an automated hospital transportation system. This approach can be used for early prototyping of PADTCs as it needs no hand-written code in the first place, but it still allows for the iterative evolvement of the application. This empowers domain experts to create their PADTC prototypes.
Multi-column dependencies in relational databases come associated with two different computational tasks. The detection problem is to decide whether a dependency of a certain type and size holds in a given database, the discovery problem asks to enumerate all valid dependencies of that type. We settle the complexity of both of these problems for unique column combinations (UCCs), functional dependencies (FDs), and inclusion dependencies (INDs). We show that the detection of UCCs and FDs is W[2]-complete when parameterized by the solution size. The discovery of inclusion-wise minimal UCCs is proven to be equivalent under parsimonious reductions to the transversal hypergraph problem of enumerating the minimal hitting sets of a hypergraph. The discovery of FDs is equivalent to the simultaneous enumeration of the hitting sets of multiple input hypergraphs. We further identify the detection of INDs as one of the first natural W[3]-complete problems. The discovery of maximal INDs is shown to be equivalent to enumerating the maximal satisfying assignments of antimonotone, 3-normalized Boolean formulas.
In order to achieve their business goals, organizations heavily rely on the operational excellence of their business processes. In traditional scenarios, business processes are usually well-structured, clearly specifying when and how certain tasks have to be executed. Flexible and knowledge-intensive processes are gathering momentum, where a knowledge worker drives the execution of a process case and determines the exact process path at runtime. In the case of an exception, the knowledge worker decides on an appropriate handling. While there is initial work on exception handling in well-structured business processes, exceptions in case management have not been sufficiently researched. This paper proposes an exception handling framework for stage-oriented case management languages, namely Guard Stage Milestone Model, Case Management Model and Notation, and Fragment-based Case Management. The effectiveness of the framework is evaluated with two real-world use cases showing that it covers all relevant exceptions and proposed handling strategies.
Effective query optimization is a core feature of any database management system. While most query optimization techniques make use of simple metadata, such as cardinalities and other basic statistics, other optimization techniques are based on more advanced metadata including data dependencies, such as functional, uniqueness, order, or inclusion dependencies. This survey provides an overview, intuitive descriptions, and classifications of query optimization and execution strategies that are enabled by data dependencies. We consider the most popular types of data dependencies and focus on optimization strategies that target the optimization of relational database queries. The survey supports database vendors to identify optimization opportunities as well as DBMS researchers to find related work and open research questions.
We consider the subset selection problem for function f with constraint bound B that changes over time. Within the area of submodular optimization, various greedy approaches are commonly used. For dynamic environments we observe that the adaptive variants of these greedy approaches are not able to maintain their approximation quality. Investigating the recently introduced POMC Pareto optimization approach, we show that this algorithm efficiently computes a phi=(alpha(f)/2)(1 - 1/e(alpha)f)-approximation, where alpha(f) is the submodularity ratio of f, for each possible constraint bound b <= B. Furthermore, we show that POMC is able to adapt its set of solutions quickly in the case that B increases. Our experimental investigations for the influence maximization in social networks show the advantage of POMC over generalized greedy algorithms. We also consider EAMC, a new evolutionary algorithm with polynomial expected time guarantee to maintain phi approximation ratio, and NSGA-II with two different population sizes as advanced multi-objective optimization algorithm, to demonstrate their challenges in optimizing the maximum coverage problem. Our empirical analysis shows that, within the same number of evaluations, POMC is able to perform as good as NSGA-II under linear constraint, while EAMC performs significantly worse than all considered algorithms in most cases.
An unceasing problem of our prevailing society is the fair division of goods. The problem of proportional cake cutting focuses on dividing a heterogeneous and divisible resource, the cake, among n players who value pieces according to their own measure function. The goal is to assign each player a not necessarily connected part of the cake that the player evaluates at least as much as her proportional share. <br /> In this article, we investigate the problem of proportional division with unequal shares, where each player is entitled to receive a predetermined portion of the cake. Our main contribution is threefold. First we present a protocol for integer demands, which delivers a proportional solution in fewer queries than all known protocols. By giving a matching lower bound, we then show that our protocol is asymptotically the fastest possible. Finally, we turn to irrational demands and solve the proportional cake cutting problem by reducing it to the same problem with integer demands only. All results remain valid in a highly general cake cutting model, which can be of independent interest.
We elaborate on the possibilities and needs to integrate design thinking into requirements engineering, drawing from our research and project experiences. We suggest three approaches for tailoring and integrating design thinking and requirements engineering with complementary synergies and point at open challenges for research and practice.
Recently, initial conflicts were introduced in the framework of M-adhesive categories as an important optimization of critical pairs. In particular, they represent a proper subset such that each conflict is represented in a minimal context by a unique initial one. The theory of critical pairs has been extended in the framework of M-adhesive categories to rules with nested application conditions (ACs), restricting the applicability of a rule and generalizing the well-known negative application conditions. A notion of initial conflicts for rules with ACs does not exist yet.
In this paper, on the one hand, we extend the theory of initial conflicts in the framework of M-adhesive categories to transformation rules with ACs. They represent a proper subset again of critical pairs for rules with ACs, and represent each conflict in a minimal context uniquely. They are moreover symbolic because we can show that in general no finite and complete set of conflicts for rules with ACs exists. On the other hand, we show that critical pairs are minimally M-complete, whereas initial conflicts are minimally complete. Finally, we introduce important special cases of rules with ACs for which we can obtain finite, minimally (M-)complete sets of conflicts.
Evaluating creativity of verbal responses or texts is a challenging task due to psychometric issues associated with subjective ratings and the peculiarities of textual data. We explore an approach to objectively assess the creativity of responses in a sentence generation task to 1) better understand what language-related aspects are valued by human raters and 2) further advance the developments toward automating creativity evaluations. Over the course of two prior studies, participants generated 989 four-word sentences based on a four-letter prompt with the instruction to be creative. We developed an algorithm that scores each sentence on eight different metrics including 1) general word infrequency, 2) word combination infrequency, 3) context-specific word uniqueness, 4) syntax uniqueness, 5) rhyme, 6) phonetic similarity, and similarity of 7) sequence spelling and 8) semantic meaning to the cue. The text metrics were then used to explain the averaged creativity ratings of eight human raters. We found six metrics to be significantly correlated with the human ratings, explaining a total of 16% of their variance. We conclude that the creative impression of sentences is partly driven by different aspects of novelty in word choice and syntax, as well as rhythm and sound, which are amenable to objective assessment.
While many optimization problems work with a fixed number of decision variables and thus a fixed-length representation of possible solutions, genetic programming (GP) works on variable-length representations. A naturally occurring problem is that of bloat, that is, the unnecessary growth of solution lengths, which may slow down the optimization process. So far, the mathematical runtime analysis could not deal well with bloat and required explicit assumptions limiting bloat.
In this paper, we provide the first mathematical runtime analysis of a GP algorithm that does not require any assumptions on the bloat. Previous performance guarantees were only proven conditionally for runs in which no strong bloat occurs. Together with improved analyses for the case with bloat restrictions our results show that such assumptions on the bloat are not necessary and that the algorithm is efficient without explicit bloat control mechanism.
More specifically, we analyzed the performance of the (1 + 1) GP on the two benchmark functions ORDER and MAJORITY. When using lexicographic parsimony pressure as bloat control, we show a tight runtime estimate of O(T-init + nlogn) iterations both for ORDER and MAJORITY. For the case without bloat control, the bounds O(T-init logT(i)(nit) + n(logn)(3)) and Omega(T-init + nlogn) (and Omega(T-init log T-init) for n = 1) hold for MAJORITY(1).
Bitcoin is gaining traction as an alternative store of value. Its market capitalization transcends all other cryptocurrencies in the market. But its high monetary value also makes it an attractive target to cyber criminal actors. Hacking campaigns usually target an ecosystem's weakest points. In Bitcoin, the exchange platforms are one of them. Each exchange breach is a threat not only to direct victims, but to the credibility of Bitcoin's entire ecosystem. Based on an extensive analysis of 36 breaches of Bitcoin exchanges, we show the attack patterns used to exploit Bitcoin exchange platforms using an industry standard for reporting intelligence on cyber security breaches. Based on this we are able to provide an overview of the most common attack vectors, showing that all except three hacks were possible due to relatively lax security. We show that while the security regimen of Bitcoin exchanges is subpar compared to other financial service providers, the use of stolen credentials, which does not require any hacking, is decreasing. We also show that the amount of BTC taken during a breach is decreasing, as well as the exchanges that terminate after being breached. Furthermore we show that overall security posture has improved, but still has major flaws. To discover adversarial methods post-breach, we have analyzed two cases of BTC laundering. Through this analysis we provide insight into how exchange platforms with lax cyber security even further increase the intermediary risk introduced by them into the Bitcoin ecosystem.
Background and objectives
AKI treated with dialysis initiation is a common complication of coronavirus disease 2019 (COVID-19) among hospitalized patients. However, dialysis supplies and personnel are often limited.
Design, setting, participants, & measurements
Using data from adult patients hospitalized with COVID-19 from five hospitals from theMount Sinai Health System who were admitted between March 10 and December 26, 2020, we developed and validated several models (logistic regression, Least Absolute Shrinkage and Selection Operator (LASSO), random forest, and eXtreme GradientBoosting [XGBoost; with and without imputation]) for predicting treatment with dialysis or death at various time horizons (1, 3, 5, and 7 days) after hospital admission. Patients admitted to theMount Sinai Hospital were used for internal validation, whereas the other hospitals formed part of the external validation cohort. Features included demographics, comorbidities, and laboratory and vital signs within 12 hours of hospital admission.
Results
A total of 6093 patients (2442 in training and 3651 in external validation) were included in the final cohort. Of the different modeling approaches used, XGBoost without imputation had the highest area under the receiver operating characteristic (AUROC) curve on internal validation (range of 0.93-0.98) and area under the precisionrecall curve (AUPRC; range of 0.78-0.82) for all time points. XGBoost without imputation also had the highest test parameters on external validation (AUROC range of 0.85-0.87, and AUPRC range of 0.27-0.54) across all time windows. XGBoost without imputation outperformed all models with higher precision and recall (mean difference in AUROC of 0.04; mean difference in AUPRC of 0.15). Features of creatinine, BUN, and red cell distribution width were major drivers of the model's prediction.
Conclusions
An XGBoost model without imputation for prediction of a composite outcome of either death or dialysis in patients positive for COVID-19 had the best performance, as compared with standard and other machine learning models.
Background:
COVID-19 has infected millions of people worldwide and is responsible for several hundred thousand fatalities. The COVID-19 pandemic has necessitated thoughtful resource allocation and early identification of high-risk patients. However, effective methods to meet these needs are lacking.
Objective:
The aims of this study were to analyze the electronic health records (EHRs) of patients who tested positive for COVID-19 and were admitted to hospitals in the Mount Sinai Health System in New York City; to develop machine learning models for making predictions about the hospital course of the patients over clinically meaningful time horizons based on patient characteristics at admission; and to assess the performance of these models at multiple hospitals and time points.
Methods:
We used Extreme Gradient Boosting (XGBoost) and baseline comparator models to predict in-hospital mortality and critical events at time windows of 3, 5, 7, and 10 days from admission. Our study population included harmonized EHR data from five hospitals in New York City for 4098 COVID-19-positive patients admitted from March 15 to May 22, 2020. The models were first trained on patients from a single hospital (n=1514) before or on May 1, externally validated on patients from four other hospitals (n=2201) before or on May 1, and prospectively validated on all patients after May 1 (n=383). Finally, we established model interpretability to identify and rank variables that drive model predictions.
Results:
Upon cross-validation, the XGBoost classifier outperformed baseline models, with an area under the receiver operating characteristic curve (AUC-ROC) for mortality of 0.89 at 3 days, 0.85 at 5 and 7 days, and 0.84 at 10 days. XGBoost also performed well for critical event prediction, with an AUC-ROC of 0.80 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. In external validation, XGBoost achieved an AUC-ROC of 0.88 at 3 days, 0.86 at 5 days, 0.86 at 7 days, and 0.84 at 10 days for mortality prediction. Similarly, the unimputed XGBoost model achieved an AUC-ROC of 0.78 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. Trends in performance on prospective validation sets were similar. At 7 days, acute kidney injury on admission, elevated LDH, tachypnea, and hyperglycemia were the strongest drivers of critical event prediction, while higher age, anion gap, and C-reactive protein were the strongest drivers of mortality prediction.
Conclusions:
We externally and prospectively trained and validated machine learning models for mortality and critical events for patients with COVID-19 at different time horizons. These models identified at-risk patients and uncovered underlying relationships that predicted outcomes.
In the stable marriage problem, a set of men and a set of women are given, each of whom has a strictly ordered preference list over the acceptable agents in the opposite class. A matching is called stable if it is not blocked by any pair of agents, who mutually prefer each other to their respective partner. Ties in the preferences allow for three different definitions for a stable matching: weak, strong and super-stability. Besides this, acceptable pairs in the instance can be restricted in their ability of blocking a matching or being part of it, which again generates three categories of restrictions on acceptable pairs. Forced pairs must be in a stable matching, forbidden pairs must not appear in it, and lastly, free pairs cannot block any matching.
Our computational complexity study targets the existence of a stable solution for each of the three stability definitions, in the presence of each of the three types of restricted pairs. We solve all cases that were still open. As a byproduct, we also derive that the maximum size weakly stable matching problem is hard even in very dense graphs, which may be of independent interest.
With the advent of big data and data lakes, data are often integrated from multiple sources. Such integrated data are often of poor quality, due to inconsistencies, errors, and so forth. One way to check the quality of data is to infer functional dependencies (fds). However, in many modern applications it might be necessary to extract properties and relationships that are not captured through fds, due to the necessity to admit exceptions, or to consider similarity rather than equality of data values. Relaxed fds (rfds) have been introduced to meet these needs, but their discovery from data adds further complexity to an already complex problem, also due to the necessity of specifying similarity and validity thresholds. We propose Domino, a new discovery algorithm for rfds that exploits the concept of dominance in order to derive similarity thresholds of attribute values while inferring rfds. An experimental evaluation on real datasets demonstrates the discovery performance and the effectiveness of the proposed algorithm.