A5020021079- Natural Language Processing Techniques
- Topic Modeling
- Text Readability and Simplification
- Data Mining Algorithms and Applications
- Linguistics and language evolution
- Machine Learning and Data Classification
- Advanced Text Analysis Techniques
- Explainable Artificial Intelligence (XAI)
- Neural Networks and Applications
- Hate Speech and Cyberbullying Detection
- Anomaly Detection Techniques and Applications
- Advanced Graph Neural Networks
- Parkinson's Disease Mechanisms and Treatments
- Rough Sets and Fuzzy Logic
- Time Series Analysis and Forecasting
- Religious, Philosophical, and Educational Studies
- Linguistics, Language Diversity, and Identity
- Biomedical Text Mining and Ontologies
- Digital Communication and Language
- Bioinformatics and Genomic Networks
- Forecasting Techniques and Applications
- Adversarial Robustness in Machine Learning
- Complex Network Analysis Techniques
- Multimodal Machine Learning Applications
- Big Data and Business Intelligence
University of Ljubljana
2015-2025
Institut za filozofiju
2016
We present a method for explaining predictions individual instances. The presented approach is general and can be used with all classification models that output probabilities. It based on decomposition of model's contributions each attribute. Our works so called black box such as support vector machines, neural networks, nearest neighbor algorithms well ensemble methods, boosting random forests. demonstrate the generated explanations closely follow learned visualization technique which...
Increasing amounts of freely available data both in textual and relational form offers exploration richer document representations, potentially improving the model performance robustness. An emerging problem modern era is fake news detection -- many easily pieces information are not necessarily factually correct, can lead to wrong conclusions or used for manipulation. In this work we explore how different ranging from simple symbolic bag-of-words, contextual, neural language model-based ones...
Abstract We present a set of novel neural supervised and unsupervised approaches for determining the readability documents. In setting, we leverage language models, whereas in three different classification architectures are tested. show that proposed approach is robust, transferable across languages, allows adaptation to specific task data set. By systematic comparison several on number benchmark new labeled sets two this study also offers comprehensive analysis classification. expose their...
Machine learning (ML) and artificial intelligence are emerging as important components of precision medicine that enhance diagnosis risk stratification. Risk stratification tools for hypertrophic cardiomyopathy (HCM) exist, but they based on traditional statistical methods. The aim was to develop a novel machine tool the prediction 5-year in HCM. goal determine if its predictive accuracy is higher than state-of-the-art tools. Data from total 2302 patients were used. data comprised...
Abstract Hate speech is an important problem in the management of user-generated content. To remove offensive content or ban misbehaving users, moderators need reliable hate detectors. Recently, deep neural networks based on transformer architecture, such as (multilingual) BERT model, have achieved superior performance many natural language classification tasks, including detection. So far, these methods not been able to quantify their output terms reliability. We propose a Bayesian method...
ABSTRACT The idle crayfish ( Austropotamobius bihariensis Pârvulescu, 2019), endemic to Romania's Apuseni Mountains, urgently requires a specific conservation plan. Due its recent description, efforts have been limited, highlighting the need for immediate and practical recommendations ensure protection. Over 13 years, field observations were conducted evaluate population trends identify threats following IUCN standards. Additionally, geospatial assessments predictive modelling employed...
There are plenty of problems where the data available is scarce and expensive. We propose a generator semi-artificial with similar properties to original which enables development testing different mining algorithms optimization their parameters. The generated allow large scale experimentation simulations without danger overfitting. proposed based on RBF networks, learn sets Gaussian kernels. These kernels can be used in generative mode generate new from same distributions. To assess quality...
Purpose The purpose of this paper is to address the problem weak acceptance machine learning (ML) models in business. proposed framework top-performing ML coupled with general explanation methods provides additional information decision-making process. This builds a foundation for sustainable organizational learning. Design/methodology/approach To user acceptance, participatory approach action design research (ADR) was chosen. demonstrated on B2B sales forecasting process an setting,...
Abstract The International Basketball Federation (FIBA) recently introduced major rule changes that came into effect with the 2010/11 season. Most notably, moving three-point arc and changing shot-clock. purpose of this study was to investigate quantify how these affect game performance top-level European basketball players. In order better understand changes, we also investigated past seasons showed presence several trends, even in absence significant changes. A large set statistics for 10...
Data preprocessing is an important component of machine learning pipelines, which requires ample time and resources. An integral part data transformation into the format required by a given algorithm. This paper outlines some modern processing techniques used in relational that enable fusion from different input types formats single table representation, focusing on propositionalization embedding approaches. While both approaches aim at transforming tabular format, they use terminology task...
Abstract Background and Purpose: The process of business to (B2B) sales forecasting is a complex decision-making process. There are many approaches support this process, but mainly it still based on the subjective judgment decision-maker. problem B2B can be modeled as classification problem. However, top performing machine learning (ML) models black boxes do not transparent reasoning. purpose research develop an organizational model using ML coupled with general explanation methods. goal...