1. The OpenScience Slovenia metadata datasetMladen Borovič, Marko Ferme, Janez Brezovnik, Sandi Majninger, Albin Bregant, Goran Hrovat, Milan Ojsteršek, 2020, drugi znanstveni članki Opis: The OpenScience Slovenia metadata dataset contains metadata entries for Slovenian public domain academic documents which include undergraduate and postgraduate theses, research and professional articles, along with other academic document types. The data within the dataset was collected as a part of the establishment of the Slovenian Open-Access Infrastructure which defined a unified document collection process and cataloguing for universities in Slovenia within the infrastructure repositories. The data was collected from several already established but separate library systems in Slovenia and merged into a single metadata scheme using metadata deduplication and merging techniques. It consists of text and numerical fields, representing attributes that describe documents. These attributes include document titles, keywords, abstracts, typologies, authors, issue years and other identifiers such as URL and UDC. The potential of this dataset lies especially in text mining and text classification tasks and can also be used in development or benchmarking of content-based recommender systems on real-world data.
Ključne besede: metadata, real world data, text data, text mining, text identification, natural language processing
Objavljeno v DKUM: 22.05.2025; Ogledov: 0; Prenosov: 8
 Celotno besedilo (187,50 KB)
Gradivo ima več datotek! Več... |
2. A brief review on benchmarking for large language models evaluation in healthcareLeona Cilar Budler, Hongyu Chen, Aokun Chen, Maxim Topaz, Wilson Tam, Jiang Bian, Gregor Štiglic, 2025, pregledni znanstveni članek Opis: This paper reviews benchmarking methods for evaluating large language models (LLMs) in healthcare settings. It highlights the importance of rigorous benchmarking to ensure LLMs' safety, accuracy, and effectiveness in clinical applications. The review also discusses the challenges of developing standardized benchmarks and metrics tailored to healthcare-specific tasks such as medical text generation, disease diagnosis, and patient management. Ethical considerations, including privacy, data security, and bias, are also addressed, underscoring the need for multidisciplinary collaboration to establish robust benchmarking frameworks that facilitate LLMs' reliable and ethical use in healthcare. Evaluation of LLMs remains challenging due to the lack of standardized healthcare-specific benchmarks and comprehensive datasets. Key concerns include patient safety, data privacy, model bias, and better explainability, all of which impact the overall trustworthiness of LLMs in clinical settings.
Ključne besede: artificial intelligence, benchmarking, chatbots, healthcare, large language models, natural language processing
Objavljeno v DKUM: 12.05.2025; Ogledov: 0; Prenosov: 3
Celotno besedilo (943,83 KB)
Gradivo ima več datotek! Več... |
3. |
4. A question answering system on domain specific knowledge with semantic web supportBorut Gorenjak, Marko Ferme, Milan Ojsteršek, 2011, izvirni znanstveni članek Opis: In today's world the majority of information is accessible via the World Wide Web. A common way to access this information is through information retrieval applications like web search engines. We already know that web search engines flood their users with enormous amount of data from which they cannot figure out the essential and most important information. These disadvantages can be reduced with question answering systems. The basic idea of question answering systems is to be able to provide answers to a specific question written in natural language. The main goal of question answering systems is to find a specific answer. This paper presents an architecture of our ontology-driven system that uses semantic description of the processes, databases and web services for question answering system in the Slovenian language.
Ključne besede: ontology, natural language processing, question answering system, semantic web, web ontology language
Objavljeno v DKUM: 01.06.2012; Ogledov: 3263; Prenosov: 77
Celotno besedilo (807,82 KB) |
5. TextProc - a natural language processing framework and its use as plagiarism detection systemJanez Brezovnik, Milan Ojsteršek, 2011, izvirni znanstveni članek Opis: A natural language processing framework called TextProc is described in this paper. First the frameworks software architecture is described. The architecture is made of several parts and all of them are described in detail. Natural language processing capabilities are implemented as software plug-ins. Plug-ins can be put together into processes that perform a practical natural processing function. Several practical TextProc processes are briefly described, like part-of-speech tagging, named entity tagging and others. One of those is capable to perform plagiarism detection on texts in Slovenian language, which is explained in detail. This process is actually used in digital library of University of Maribor. The integration of digital library with TextProc is also briefly described. At the end of this paper some ideas for future development are given.
Ključne besede: natural language processing, text processing, text mining, Slovenian language, plagiarism detection
Objavljeno v DKUM: 01.06.2012; Ogledov: 3022; Prenosov: 95
Celotno besedilo (438,30 KB) |
6. Advanced features of Digital library of University of MariborJanez Brezovnik, Milan Ojsteršek, 2011, izvirni znanstveni članek Opis: Advanced features of digital library of University of Maribor are described inthis paper. A short introduction describes some basic facts about the digital library and mentions its main purpose, but the main part of this paperis about features, that are mostly not found in other digital libraries. These features include integration with other information systems, plagiarism detection, informative and useful statistics about mentors and specific content extraction from documents, served by the digital library. We present existing functionality and describe some ideas for future development. A natural language processing framework, called TextProc, is also briefly mentioned, since it is used to perform plagiarism detection.
Ključne besede: digital library, natural language, plagiarism detection, Slovenian language, text processing
Objavljeno v DKUM: 01.06.2012; Ogledov: 3167; Prenosov: 65
 Povezava na celotno besedilo |
7. Local search engine with global content based on domain specific knowledgeSandi Pohorec, Mateja Verlič, Milan Zorman, 2009, izvirni znanstveni članek Opis: In the growing need for information we have come to rely on search engines. The use of large scale search engines, such as Google, is as common as surfingthe World Wide Web. We are impressed with the capabilities of these search engines but still there is a need for improvment. A common problem withsearching is the ambiguity of words. Their meaning often depends on the context in which they are used or varies across specific domains. To resolve this we propose a domain specific search engine that is globally oriented. We intend to provide content classification according to the target domain concepts, access to privileged information, personalization and custom rankingfunctions. Domain specific concepts have been formalized in the form ofontology. The paper describes our approach to a centralized search service for domain specific content. The approach uses automated indexing for various content sources that can be found in the form of a relational database, we! b service, web portal or page, various document formats and other structured or unstructured data. The gathered data is tagged with various approaches and classified against the domain classification. The indexed data is accessible through a highly optimized and personalized search service.
Ključne besede: information search, personalization, indexes, crawling, domain specific crawling, natural language processing, content tagging, distributed data sources, ranking functions
Objavljeno v DKUM: 31.05.2012; Ogledov: 2151; Prenosov: 37
Povezava na celotno besedilo |
