- Fuzzy Logic and Control Systems
- Anomaly Detection Techniques and Applications
- Machine Learning in Healthcare
- Neural Networks and Applications
- Time Series Analysis and Forecasting
- Music Technology and Sound Studies
- Target Tracking and Data Fusion in Sensor Networks
- Explainable Artificial Intelligence (XAI)
- Advanced Optical Sensing Technologies
- Structural Health Monitoring Techniques
- Adversarial Robustness in Machine Learning
- Advanced Chemical Sensor Technologies
- Artificial Intelligence in Healthcare
- Air Quality Monitoring and Forecasting
- Biomedical Text Mining and Ontologies
- Music and Audio Processing
- Petri Nets in System Modeling
- Intelligent Tutoring Systems and Adaptive Learning
- Privacy-Preserving Technologies in Data
- Non-Invasive Vital Sign Monitoring
- Respiratory Support and Mechanisms
- Gene Regulatory Network Analysis
- Topic Modeling
- Advanced Graph Neural Networks
- COVID-19 diagnosis using AI
Trier University of Applied Sciences
2019-2024
University of Sarajevo
2009-2017
Abstract The aim of this work was to develop and evaluate the reinforcement learning algorithm VentAI, which is able suggest a dynamically optimized mechanical ventilation regime for critically-ill patients. We built, validated tested its performance on 11,943 events volume-controlled derived from 61,532 distinct ICU admissions it an independent, secondary dataset (200,859 stays; 25,086 events). A patient “data fingerprint” 44 features extracted as multidimensional time series in 4-hour...
Medical guidelines have a significant role in the field of evidence-based medical treatment. The content guideline is based on systematic review clinical evidence with instructions and recommendations that clinicians can refer to. Most are available an unstructured text format. Hence, must take considerable time to search find relevant their semantic context. Using Machine Learning algorithms, automatic information extraction from has recently become possible. We present novel system for...
Background High numbers of consumable medical materials (eg, sterile needles and swabs) are used during the daily routine intensive care units (ICUs) worldwide. Although consumables largely contribute to total ICU hospital expenditure, many hospitals do not track individual use materials. Current tracking solutions meeting specific requirements environment, like barcodes or radio frequency identification, require specialized material preparation high infrastructure investment. This impedes...
This paper presents novel GA-ANFIS expert system prototype for dermatological disease detection by using features and diagnoses collected in real conditions. Nine are used as inputs to classifiers that based on Adaptive Neuro-Fuzzy Inference Systems (ANFIS) the first level of fuzzy model optimization. After that, they Genetic Algorithm (GA) second optimization within system. performs two steps. Modelling validation approach is performed MATLAB environment set data. Some conclusions...
In recent years, the volume of medical knowledge and health data has increased rapidly. For example, availability electronic records (EHRs) provides accurate, up-to-date, complete information about patients at point care enables staff to have quick access patient for more coordinated efficient care. With this increase in knowledge, complexity evidence-based medicine tends grow all time. Health workers must deal with an increasing amount documentation. Meanwhile, relevant are frequently...
The purpose of this study is to present novel GAANFIS expert system prototype for tar detection in cigarettes during manufacturing process. proposed combines capabilities Adaptive Neuro-Fuzzy Inference System (ANFIS) and Genetic Algorithm (GA).The data recorded different type are collected by special control quality equipment real conditions inside cigarette factory. GA-ANFIS performs optimization two steps. In the first step it generates six ANFIS structures, after that, we have second...
Abstract 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 sepsis one such case. Early treatment cost effective reduces number hospital days patients ICU. There no single procedure that sufficient for diagnosis, combinations approaches needed. Detecting patient time series data could help speed development some decisions. However, our...
Deep Learning (DL) is being applied in various domains, especially safety-critical applications such as autonomous driving. Consequently, it of great significance to ensure the robustness these methods and thus counteract uncertain behaviors caused by adversarial attacks. In this paper, we use gradient heatmaps analyze response characteristics VGG-16 model when input images are mixed with noise statistically similar Gaussian random noise. particular, compare network layer determine where...
In this work, a new hybrid algorithm for disease risk classification is proposed. The proposed methodology based on Dynamic Time Warping (DTW). This can be applied to time series from various domains such as vital sign available in medical big data. To validate our methodology, we it sepsis, which one of the most challenging problems within area data analysis. first step uses different statistical properties features. Furthermore, using differently labeled training sets, created DTW...
In this paper, we propose an innovative method for determining the fill level of containers, such as trash cans, addressing a critical aspect waste management. The combines spatial impulse response analysis with machine learning techniques, offering unique and effective approach sound-based classification that can be extended to various domains beyond By employing buzzer-generated sine sweep signal, create distinctive signature specific container. This is then interpreted by specially...
This paper presents an incremental learning approach for estimating the structural parameters in stochastic state-based models (SSMs). SSMs have proven to be useful modelling biological and medical processes, as they can represent both time dependency processes. A major challenge bioinformatics is that processes usually rely on large publicly accessible databases. In this work, a new presented, where are trained incrementally locally at different data sources, e.g., hospitals, without having...
The purpose of this study is to model and optimize the detection tar in cigarettes during manufacturing process show that low yield contain similar levels nicotine as compared high while B (Benzene), T(toluene) X (xylene) (BTX) increase with increasing yields. A neuro-fuzzy system which comprises a fuzzy inference structure used such system. Given training set samples, Adaptive Neuro-Fuzzy Inference System (ANFIS) classifiers learned how differentiate new case domain. ANFIS were detect smoke...
This paper presents novel GA-ANFIS expert system approach for the prediction of lung nicotine concentration by using "on inspect machine" recorded data. Smoking causes majority cancers — both in smokers and people exposed to secondhand smoke. Total yields cigarette smoke constituents are greatly influenced smoking behavior. The size particles inhaled directly from a is controlled ventilation it has largest influence on reducing tar, carbon monoxide yields. Our goal get better results as well...
The aim of this research is to develop a novel GA-ANFIS expert system prototype for classifying heart disease degree patient by using diseases attributes (features) and diagnoses taken in the real conditions. Thirteen have been used as inputs classifiers being based on Adaptive Neuro-Fuzzy Inference Systems (ANFIS) first level fuzzy model optimization. They are Genetic Algorithm (GA) second optimization within system. performs two steps. Modelling validating approach performed MATLAB...
A study is presented for the detection of nicotine impact in different cigarette type, using recorded data and Computational Intelligence techniques. Recorded puffs are processed Continuous Wavelet Transform used to extract time-frequency features normal abnormal conditions. The wavelet energy distributions as inputs classifiers based on Adaptive Neuro-Fuzzy Inference Systems (ANFIS) Genetic Algorithms (GAs). number parameters Membership Functions ANFIS along with from distributionare...
Deep Learning (DL) is being applied in various domains, especially safety-critical applications such as autonomous driving. Consequently, it of great significance to ensure the robustness these methods and thus counteract uncertain behaviors caused by adversarial attacks. In this paper, we use gradient heatmaps analyze response characteristics VGG-16 model when input images are mixed with noise statistically similar Gaussian random noise. particular, compare network layer determine where...
An important task in the field of sensor technology is efficient implementation adaptation procedures measurements from one to another identical design. One idea use estimation an affine transformation between different systems, which can be improved by knowledge experts. This paper presents solution Glacier Research that was published back 1973. The results demonstrate adaptability this for various applications, including software calibration sensors, expert-based adaptation, and paving way...
An important task in the field of sensor technology is efficient implementation adaptation procedures measurements from one to another identical design.One idea use estimation an affine transformation between different systems, which can be improved by knowledge experts.This paper presents solution Glacier Research that was published back 1973.The results demonstrate adaptability this for various applications, including software calibration sensors, expert-based adaptation, and paving way...
The aim of this study is to present novel algorithms for prediction dermatological disease using only clinical features and diagnoses collected in real conditions. A combination the Adaptive Neuro-Fuzzy Inference Systems (ANFIS) Genetic algorithm (GA) ANFIS subtractive clustering parameter optimization has been suggested first level fuzzy model optimization. After that, a genetic optimized structure used as input GA second We double 2-fold Cross validation generating different sets...
Data analysis and their application are the unavoidable factors in activities analyses health care. Unfortunately, acquisition of data from large available medical databases is a complex process requires deep knowledge computer science especially tools for management. According to European General Protection Regulation, problem becomes much more complex. Recognizing these problems difficulties, we have developed Science Learning Platform (DSLP) that primarily targets practitioners...
The availability of Big Data has increased significantly in many areas recent years. Insights from these data sets lead to optimized processes industries, which is why understanding as well gaining knowledge through analyses becoming increasingly relevant. In the medical field, especially intensive care units, fast and appropriate treatment crucial due usually critical condition patients. patient recorded here often very heterogeneous resulting database models are complex, so that accessing...
Window ventilation is important in everyday life. The COVID-19 pandemic particular has shown that air exchange necessary to minimize the spread of viruses. Efficient can be supported with help sensors and intelligent data processing. A CO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> sensor, for example, used measure levels and, together IoT hardware, indicating when needed. By combining these components algorithms, an assessment such...