A5045105508- Gene expression and cancer classification
- Bayesian Modeling and Causal Inference
- Neural Networks and Applications
- Machine Learning and Data Classification
- Machine Learning in Bioinformatics
- Fractal and DNA sequence analysis
- Parkinson's Disease Mechanisms and Treatments
- Human Mobility and Location-Based Analysis
- Amyotrophic Lateral Sclerosis Research
- Anomaly Detection Techniques and Applications
- Genomic variations and chromosomal abnormalities
- Data Quality and Management
- Advanced Image and Video Retrieval Techniques
- Evolutionary Algorithms and Applications
- Face and Expression Recognition
- Algorithms and Data Compression
- Blind Source Separation Techniques
- Gaussian Processes and Bayesian Inference
- Geographic Information Systems Studies
- Fuzzy Logic and Control Systems
- Image Retrieval and Classification Techniques
- Domain Adaptation and Few-Shot Learning
- Advanced Neural Network Applications
- Explainable Artificial Intelligence (XAI)
- Neurogenetic and Muscular Disorders Research
Ben-Gurion University of the Negev
2016-2025
Hebrew University of Jerusalem
2020
Rabin Medical Center
2010
University of Cambridge
1998-2001
Bridge University
2000
Universidade da Coruña
1994
Within the complex and competitive semiconductor manufacturing industry, lot cycle time (CT) remains one of key performance indicators. Its reduction is strategic importance as it contributes to cost decreasing, time-to-market shortening, faster fault detection, achieving throughput targets, improving production-resource scheduling. To reduce CT, we suggest investigate a data-driven approach that identifies factors predicts their impact on CT. In our novel approach, first identify most...
The application of neural networks (NNs) to automatic analysis chromosome images is investigated in this paper. All aspects the analysis, namely segmentation, feature description, selection and extraction, classification, are studied. As part segmentation process, separation clusters partially occluded chromosomes, which critical stage that state-of-the-art analyzers usually fail accomplish, performed. First, a moment representation image pixels clustered create binary without need for...
In this paper, we modify the fuzzy ARTMAP (FA) neural network (NN) using Bayesian framework in order to improve its classification accuracy while simultaneously reduce category proliferation. The proposed algorithm, called (BA), preserves FA advantages and also enhances performance by following: (1) representing a multidimensional Gaussian distribution, (2) allowing grow or shrink, (3) limiting hypervolume, (4) Bayes' decision theory for learning inference, (5) employing probabilistic...
Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease where substantial heterogeneity in clinical presentation urgently requires better stratification of patients for the development drug trials and care. In this study we explored through crowdsourcing approach, DREAM Prize4Life ALS Stratification Challenge. Using data from >10,000 1479 community-based patient registers, more than 30 teams developed new approaches machine learning clustering, outperforming best current...
Abstract Background Various biomarkers may help predict inflammatory bowel disease (IBD) in advance of diagnosis, but these are mostly not routinely tested and thus more challenging for screening. In this nationwide study, we used the epi-IIRN validated cohort to explore utility routine blood tests as markers predicting IBD occurrence. Methods We collect results all performed up 15 years before from patients diagnosed between 2005-2020 insured by Israeli health maintenance organizations...
During the process of rehabilitation after stroke, it is important that patients know how well they perform their exercise, so can improve performance in future repetitions. Standard clinical rating conducted by human observation prevailing way today to monitor motor recovery patient. Therefore, cannot whether are performing a movement properly while exercising themselves. Adhering exercise regime makes more effective and efficient, thus system give feedback on great value. Here, we built...
Background: The role of the lipidome as a biomarker for Parkinson’s disease (PD) is relatively new field that currently only focuses on PD diagnosis. Objective: To identify relevant signature severity markers. Methods: Disease 149 patients was assessed by Unified Rating Scale (UPDRS) and Montreal Cognitive Assessment (MoCA). lipid composition whole blood samples analyzed, consisting 517 species from 37 classes; these included all major classes glycerophospholipids, sphingolipids,...
Background Under- or late identification of pulmonary embolism (PE)—a thrombosis 1 more arteries that seriously threatens patients’ lives—is a major challenge confronting modern medicine. Objective We aimed to establish accurate and informative machine learning (ML) models identify patients at high risk for PE as they are admitted the hospital, before their initial clinical checkup, by using only information in medical records. Methods collected demographics, comorbidities, medications data...
In many applications, the detection of a visually obscured magnetic target is followed by characterization target, i.e. localization and moment estimation. Effective may reduce false alarm rate as well direct searcher toward target. We address static three-axis fluxgate magnetometer installed on stabilized mobile platform. The readings are contaminated noise, which results in low signal-to-noise ratio. formulate problem an over-determined nonlinear equation set using dipole model for use...
We experimentally study the K2 algorithm in learning a Bayesian network (BN) classifier for image detection of cytogenetic abnormalities. Starting from an initial BN structure, searches structure space and selects maximizing metric. To improve accuracy K2-based classifier, we investigate ordering, search procedure, find that structures learned using random orderings, orderings based on expert knowledge, or scatter criterion are comparable lead to similar classification accuracies. Replacing...
Two feature selection techniques and a multilayer perceptron (MLP) neural network (NN) have been used in this study for human chromosome classification. The first technique is the "knock-out" algorithm second principal component analysis (PCA). emphasized significance of centrometric index length, as features PCA demonstrated importance retaining most image information whenever small training sets are used. However, use large enables considerable data compression. Both yield benefit using...
We propose and investigate the fuzzy ARTMAP neural network in off online classification of fluorescence situ hybridization image signals enabling clinical diagnosis numerical genetic abnormalities. evaluate task (detecting a several abnormalities separately or simultaneously), classifier paradigm (monolithic hierarchical), ordering strategy for training patterns (averaging voting), mode (for one epoch, with validation until completion) model sensitivity to parameters. find accurate...
Objective: Amyotrophic lateral sclerosis (ALS) disease state prediction usually assumes linear progression and uses a classifier evaluated by its accuracy. Since is not linear, the accuracy measurement cannot tell large from small errors, we dispense with linearity assumption apply ordinal classification that accounts for error severity. In addition, identify most influential variables in predicting explaining disease. Furthermore, contrast to conventional modeling of patient's total...
Classification of segment images created by connecting points high concavity along curvatures is used to resolve partial occlusion in images. Modeling shape or curvature not necessary nor the traditional excessive use heuristics. Applied human cell images, 82.6% analyzed clusters chromosomes are correctly separated, rising 90.5% following rejection 8.7%
Abstract Lipid profiles in biological fluids from patients with Parkinson’s disease (PD) are increasingly investigated search of biomarkers. However, the lipid genetic PD remain to be determined, a gap knowledge particular interest associated mutant α-synuclein ( SNCA ), given known relationship between this protein and lipids. The objective research is identify serum composition A53T mutation carriers compare these alterations those found cells transgenic mice carrying same mutation. We...
Background: Previous systems for dot (signal) counting in fluorescence situ hybridization (FISH) images have relied on an auto-focusing method obtaining a clearly defined image. Because signals are distributed three dimensions within the nucleus and artifacts such as debris background can attract focusing method, valid be left unfocused or unseen. This leads to errors, which increase with number of probes. Methods: The approach described here dispenses auto-focusing, instead relies neural...