00 Informatik, Wissen, Systeme
Filtern
Dokumenttyp
Sprache
- Englisch (18) (entfernen)
Volltext vorhanden
- ja (18)
Gehört zur Bibliographie
- nein (18) (entfernen)
Schlagworte
- Maschinelles Lernen (5)
- machine learning (4)
- Convolutional Neural Network (3)
- Künstliche Intelligenz (3)
- Alkoholmissbrauch (2)
- Anomalie <Medizin> (2)
- Anomalieerkennung (2)
- App <Programm> (2)
- Intervention <Medizin> (2)
- Jugend (2)
Online Learning algorithms and Indoor Positioning Systems are complex applications in the environment of cyber-physical systems. These distributed systems are created by networking intelligent machines and autonomous robots on the Internet of Things using embedded systems that enable the exchange of information at any time. This information is processed by Machine Learning algorithms to make decisions about current developments in production or to influence logistics processes for optimization purposes. In this article, we present and categorize the further development of the prototype of a novel Indoor Positioning System, which constantly adapts its knowledge to the conditions of its environment with the help of Online Learning. Here, we apply Online Learning algorithms in the field of sound-based indoor localization with low-cost hardware and demonstrate the improvement of the system over its predecessor and its adaptability for different applications in an experimental case study.
Internet of Things (IoT) and Artificial Intelligence (AI) are one of the most promising and disruptive areas of current research and development. However, these areas require deep knowledge in multiple disciplines such as sensors, protocols, embedded programming, distributed systems, statistics and algorithms. This broad knowledge is not easy to acquire and the software used to design these systems is becoming increasingly complex. Small and medium-sized enterprises therefore have problems in developing new business ideas. However, node- and block-based software tools have also been released and are freely available as open source toolboxes. In this paper, we present an overview of multiple node- and block-based software tools to develop IoT- and AI-based business ideas. We arrange these tools according their capabilities and further propose extension and combinations of tools to design a useful open-source library for small and medium-sized enterprises, that is easy to use and helps with rapid prototyping, enabling new business ideas to be developed using distributed computing.
The objective investigation of the dynamic properties of vocal fold vibrations demands the recording and further quantitative analysis of laryngeal high-speed video (HSV). Quantification of the vocal fold vibration patterns requires as a first step the segmentation of the glottal area within each video frame from which the vibrating edges of the vocal folds are usually derived. Consequently, the outcome of any further vibration analysis depends on the quality of this initial segmentation process. In this work we propose for the first time a procedure to fully automatically segment not only the time-varying glottal area but also the vocal fold tissue directly from laryngeal high-speed video (HSV) using a deep Convolutional Neural Network (CNN) approach. Eighteen different Convolutional Neural Network (CNN) network configurations were trained and evaluated on totally 13,000 high-speed video (HSV) frames obtained from 56 healthy and 74 pathologic subjects. The segmentation quality of the best performing Convolutional Neural Network (CNN) model, which uses Long Short-Term Memory (LSTM) cells to take also the temporal context into account, was intensely investigated on 15 test video sequences comprising 100 consecutive images each. As performance measures the Dice Coefficient (DC) as well as the precisions of four anatomical landmark positions were used. Over all test data a mean Dice Coefficient (DC) of 0.85 was obtained for the glottis and 0.91 and 0.90 for the right and left vocal fold (VF) respectively. The grand average precision of the identified landmarks amounts 2.2 pixels and is in the same range as comparable manual expert segmentations which can be regarded as Gold Standard. The method proposed here requires no user interaction and overcomes the limitations of current semiautomatic or computational expensive approaches. Thus, it allows also for the analysis of long high-speed video (HSV)-sequences and holds the promise to facilitate the objective analysis of vocal fold vibrations in clinical routine. The here used dataset including the ground truth will be provided freely for all scientific groups to allow a quantitative benchmarking of segmentation approaches in future.
The aim of this work was to develop and evaluate the reinforcement learning algorithm VentAI, which is able to suggest a dynamically optimized mechanical ventilation regime for critically-ill patients. We built, validated and tested its performance on 11,943 events of volume-controlled mechanical ventilation derived from 61,532 distinct ICU admissions and tested it on an independent, secondary dataset (200,859 ICU stays; 25,086 mechanical ventilation events). A patient “data fingerprint” of 44 features was extracted as multidimensional time series in 4-hour time steps. We used a Markov decision process, including a reward system and a Q-learning approach, to find the optimized settings for positive end-expiratory pressure (PEEP), fraction of inspired oxygen (FiO2) and ideal body weight-adjusted tidal volume (Vt). The observed outcome was in-hospital or 90-day mortality. VentAI reached a significantly increased estimated performance return of 83.3 (primary dataset) and 84.1 (secondary dataset) compared to physicians’ standard clinical care (51.1). The number of recommended action changes per mechanically ventilated patient constantly exceeded those of the clinicians. VentAI chose 202.9% more frequently ventilation regimes with lower Vt (5–7.5 mL/kg), but 50.8% less for regimes with higher Vt (7.5–10 mL/kg). VentAI recommended 29.3% more frequently PEEP levels of 5–7 cm H2O and 53.6% more frequently PEEP levels of 7–9 cmH2O. VentAI avoided high (>55%) FiO2 values (59.8% decrease), while preferring the range of 50–55% (140.3% increase). In conclusion, VentAI provides reproducible high performance by dynamically choosing an optimized, individualized ventilation strategy and thus might be of benefit for critically ill patients.
Optimal mental workload plays a key role in driving performance. Thus, driver-assisting systems that automatically adapt to a drivers current mental workload via brain–computer interfacing might greatly contribute to traffic safety. To design economic brain computer interfaces that do not compromise driver comfort, it is necessary to identify brain areas that are most sensitive to mental workload changes. In this study, we used functional near-infrared spectroscopy and subjective ratings to measure mental workload in two virtual driving environments with distinct demands. We found that demanding city environments induced both higher subjective workload ratings as well as higher bilateral middle frontal gyrus activation than less demanding country environments. A further analysis with higher spatial resolution revealed a center of activation in the right anterior dorsolateral prefrontal cortex. The area is highly involved in spatial working memory processing. Thus, a main component of drivers’ mental workload in complex surroundings might stem from the fact that large amounts of spatial information about the course of the road as well as other road users has to constantly be upheld, processed and updated. We propose that the right middle frontal gyrus might be a suitable region for the application of powerful small-area brain computer interfaces.
Modeling and executing knowledge-intensive processes (KiPs) are challenging with state-of-the-art approaches, and the specific demands of KiPs are the subject of ongoing research. In this context, little attention has been paid to the ontology-driven combination of data-centric and semantic business process modeling, which finds additional motivation by enabling the division of labor between humans and artificial intelligence. Such approaches have characteristics that could allow support for KiPs based on the inferencing capabilities of reasoners. We confirm this as we show that reasoners can infer the executability of tasks based on a currently researched ontology- and data-driven business process model (ODD-BP model). Further support for KiPs by the proposed inference mechanism results from its ability to infer the relevance of tasks, depending on the extent to which their execution would contribute to process progress. Besides these contributions along with the execution perspective (start-to-end direction), we will also show how our approach can help to reach specific process goals by inferring the relevance of process elements regarding their support to achieve such goals (end-to-start direction). The elements with the most valuable process progress can be identified in the intersection of both, the execution and goal perspective. This paper will introduce this new approach and verifies its practicability with an evaluation of a KiP in the field of emergency call centers.
Numerous research methods have been developed to detect anomalies in the areas of security and risk analysis. In healthcare, there are numerous use cases where anomaly detection is relevant. For example, early detection of sepsis is one such use case. Early treatment of sepsis is cost effective and reduces the number of hospital days of patients in the ICU. There is no single procedure that is sufficient for sepsis diagnosis, and combinations of approaches are needed. Detecting anomalies in patient time series data could help speed the development of some decisions. However, our algorithm must be viewed as complementary to other approaches based on laboratory values and physician judgments. The focus of this work is to develop a hybrid method for detecting anomalies that occur, for example, in multidimensional medical signals, sensor signals, or other time series in business and nature. The novelty of our approach lies in the extension and combination of existing approaches: Statistics, Self Organizing Maps and Linear Discriminant Analysis in a unique and unprecedented way with the goal of identifying different types of anomalies in real-time measurement data and defining the point where the anomaly occurs. The proposed algorithm not only has the full potential to detect anomalies, but also to find real points where an anomaly starts.
Background: In recent years, the volume of medical knowledge and health data has increased rapidly. For example, the increased availability of electronic health records (EHRs) provides accurate, up-to-date, and complete information about patients at the point of care and enables medical staff to have quick access to patient records for more coordinated and efficient care. With this increase in knowledge, the complexity of accurate, evidence-based medicine tends to grow all the time. Health care workers must deal with an increasing amount of data and documentation. Meanwhile, relevant patient data are frequently overshadowed by a layer of less relevant data, causing medical staff to often miss important values or abnormal trends and their importance to the progression of the patient’s case.
Objective: The goal of this work is to analyze the current laboratory results for patients in the intensive care unit (ICU) and classify which of these lab values could be abnormal the next time the test is done. Detecting near-future abnormalities can be useful to support clinicians in their decision-making process in the ICU by drawing their attention to the important values and focus on future lab testing, saving them both time and money. Additionally, it will give doctors more time to spend with patients, rather than skimming through a long list of lab values.
Methods: We used Structured Query Language to extract 25 lab values for mechanically ventilated patients in the ICU from the MIMIC-III and eICU data sets. Additionally, we applied time-windowed sampling and holding, and a support vector machine to fill in the missing values in the sparse time series, as well as the Tukey range to detect and delete anomalies. Then, we used the data to train 4 deep learning models for time series classification, as well as a gradient boosting–based algorithm and compared their performance on both data sets.
Results: The models tested in this work (deep neural networks and gradient boosting), combined with the preprocessing pipeline, achieved an accuracy of at least 80% on the multilabel classification task. Moreover, the model based on the multiple convolutional neural network outperformed the other algorithms on both data sets, with the accuracy exceeding 89%.
Conclusions: In this work, we show that using machine learning and deep neural networks to predict near-future abnormalities in lab values can achieve satisfactory results. Our system was trained, validated, and tested on 2 well-known data sets to ensure that our system bridged the reality gap as much as possible. Finally, the model can be used in combination with our preprocessing pipeline on real-life EHRs to improve patients’ diagnosis and treatment.