Group for language technologies
The Language Technologies Group develops and maintains resources and tools for processing the Serbian language. Its approaches combine linguistic knowledge with statistical analysis in order to address the specific characteristics of the language and its application domains as effectively and reliably as possible. Among the available resources, particularly noteworthy are the electronic morphological dictionary (approximately 3.5 million word forms), a corpus of contemporary Serbian (approximately 122 million words), and Serbian WordNet (approximately 21,000 synsets), one of the best-known semantic networks. The group also maintains English–Serbian and French–Serbian parallel corpora, and efforts are underway to create the first syntactic treebank. The group is actively involved in named entity recognition and information extraction across various domains, sentiment analysis, topic classification, information retrieval and search, and machine translation.
In October 2024, the group celebrated its 45th anniversary. During this period, numerous seminars have been organized, many important conferences have been held, and projects of international significance have been carried out. The group actively collaborates with colleagues from the Faculty of Philology, the Faculty of Mining and Geology, and the Faculty of Agriculture of the University of Belgrade, with whom it founded the Society for Language Technologies — JeRTeh — in 2014.
Further information about the work of the group and the JeRTeh Society is available at: