EN BG

Semantic Network with a Wide Range of Semantic Relations



Period: 2017-2019

Type of project: collective

Funding: National Science Fund, Grant Agreement ДН 10/3 from 14/12/2016



Principal Investigator: Prof. S. Koeva, Ph.D.

Participants: Prof. S. Koeva, Assist. Prof. S. Leseva, Assist. Prof. T. Dimitrova, Assist. Prof. M. Todorova, Assist. Prof. Valentina Stefanova, I. Stoyanova, B. Rizov, D. Hristov, M. Yalamov; prof. Tinko Tinchev (Sofia University), prof. Maciej Piasecki (Wrocław University of Science and Technology).

Abstract:

The project offers fundamental scientific research in the field of natural language semantics. Its main objective is to enrich the semantic networks with new semantic relations, which have not yet been the focus of either lexical semantics, nor known semantic networks. The Bulgarian WordNet (Koeva 2014), a lexical-semantic network for Bulgarian, will be used as a starting point for research. The enrichment is carried out throught the implementation of conceptual frames which in general encode the relation between a predicate and its arguments.

In order to define potential arguments to a given predicate, a detailed ontological description of noun and verb semantic classes within WordNet is required. There are 82,114 noun synsets in WordNet grouped into 25 semantic classes assigned to 253 semantic categories using a total of 171,359 mappings (results). 13 465 verb synsets, grouped into 15 semantic classes, are further classified into categories defined by FrameNet frames and VerbNet superclasses.

Conceptual frames are defined as abstract structures described by means of a unique set of semantic relations between: (a) a frame represented by verb predicates arranged in WordNet synsets and belonging to a particular semantic class that reflects the properties of the frame; (b) frame elements represented by noun synsets belonging to a particular semantic class that describes the properties of the frame elements; (c) semantic relations between the frame and each of its elements.

The components of the theoretical description of conceptual frames representing the combinatorial potential between a semantic relation, a verb synset from а given semantic class (or classes) and a set of noun synsets belonging to a particular semantic class (or classes) include the definition of a hierarchy of frame elements, a hierarchy of semantic (predicate-argument or predicate-adjunct) relations and the basic semantic types of the frame elements. 5,025 of the frames assigned to verb synsets in WordNet have been validated by experts.

The team has studied the conditions under which pairs of noun synsets that are not related by means of a semantic relation in WordNet may be connected semantically on the basis of existing relations with verb or adjective synsets. 21 new semantic relations have been defined and 2,814 noun synsets have been related to each other by experts through 589 relations.

2,904 patterns were automatically assigned to 2,593 verb synsets in WordNet (verb patterns consist of ordered lists of frames whose elements are realised by noun phrases sharing a common semantic component represented as the element’s semantic type). As a result of the manual expert verification and enrichment the total number of verb synsets with assigned patterns in WordNet has totalled 3,986.

The resulting semantic network enriched with a variety of semantic relations is extremely useful for scientists, editors, translators and anyone else interested in the Bulgarian language and has many applications in natural language processing, in particular in the field of: machine translation, automatic semantic analysis (opinion mining, event detection, event tracking and event prediction, text transformation and text simplification), information search and information retrieval (text summarisation, document classification, question answering).

The results obtained in stage 2 of the project were presented in 19 papers in indexed and refereed publications (SCOPUS, Web ISI, ERIH+) and at 5 international conferences. A special reviewed volume of papers was published in English and was subsequently approved for distribution by the CEEOL database and the Russian Science Citation Index (RSCI/РИНЦ). Part of the results were presented at 2 open seminars and at the Special Session on WordNet and Ontologies during the fourth edition of the International Conference Computational Linguistics in Bulgaria (CLIB 2018, CLIB 2020)

The results obtained in the project are presented in a collection of studies:
➥ Koeva, S. (Ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations. Prof. Marin Drinov Academic Publishing House of Bulgarian Academy of Sciences. Sofia, 2020. ISBN 978-619-245-057-1. DOI: 10.7546/TSN.2020


Work package 2: An analytical overview of existing research


An overview of the representation of knowledge by means of semantic networks

➥ Leseva, Svetlozara. Knowledge representation in semantic networks. / Лесева, Светлозара. Представяне на знания чрез семантични мрежи. – сп. „Български език“, кн. 2, 2018.


Research on existing descriptions of semantic relations

➥ Maciej Piasecki, Svetla Koeva. WordNet Relations in the Bulgarian-Polish Bilingual Perspective. В: Доклади от Международната юбилейна конференция на Института за български език, 2017, част I. pdf

➥ Todorova, Maria, Stoyanova, Ivelina Semantic relations: theoretical and practical aspects. / Тодорова, Мария, Ивелина Стоянова. Семантични релации: теоретични и приложни аспекти. – сп. „Български език“, кн. 2, 2018.


Research on existing semantic classifications within the following parts of speech: verbs, nouns.

➥ Dimitrova, Tsvetana. Morphosemantic relations and agentive nouns in the Bulgarian WordNet. / Димитрова, Цветана. Морфосемантични релации и агентивни съществителни в Българския Уърднет. – В: сп. „Български език“, кн. 2, 2018, с. 41-58. ISSN 0005-4283. (ERIH+) pdf


Work package 3: Specification of the semantic classes in WordNet


A detailed ontological representation of semantic classes (defined by the semantic primitives) of nouns in WordNet as determined by the semantic restrictions imposed on the noun argument in the verb – noun semantic relations

➥ WordNet with alignd CPA semantic types to noun synsets (resource)

➥ Svetla Koeva, Tsvetana Dimitrova, Valentina Stefanova, Dimitar Hristov. Mapping WordNet Concepts with CPA Ontology. In: Proceedings of GWC 2018. pdf


A detailed ontological representation of semantic classes (defined by semantic primitives) of verbs in WordNet as determined by the semantic restrictions imposed by the verb – noun semantic relations

➥ Classification of verb synsets in WordNet (resource)

➥ Svetlozara Leseva, Ivelina Stoyanova, Maria Todorova. Classifying Verbs in WordNet by Harnessing Semantic Resources. In: Proceedings of CLIB 2018. pdf


Description of the inheritance of semantic primitives between synsets containing derivationally related nouns and verbs

➥ Ivelina Stoyanova. Factors and Features Determining the Inheritance of Semantic Primes between Verbs and Nouns within WordNet. In: Proceedings of CLIB 2018. pdf


Work package 4: Definition of semantic relations that have not been encoded in WordNet


Definition of more specific semantic relations that are subsumed under certain existing semantic relations in WordNet (hyponymy, hypernymy, meronymy, holonymy, antonymy, is subevent, has subevent, causes, is caused by)

➥ Koeva, Svetla, Valentina Stefanova, Dimitar Hristov. Semantic Relations Within Multiple Hypernymy in WordNet. / Коева, Св., Стефанова, В., Христов Д. Семантични релации в рамките на многократната хиперонимия в Уърднет. – сп. „Чуждоезиково обучение“, кн. 4, 2018.


Description of new morphosemantic relations between verb and noun synsets (semantic relations that are signalled by derivational relations) in WordNet

➥ Dimitrova, Tsvetana. Morphosemantic relations and agentive nouns in the Bulgarian WordNet. / Цветана Димитрова. Морфосемантични релации и агентивни съществителни в Българския Уърднет. – сп. „Български език“, кн. 2, 2018.


Description of new semantic relations between verb and noun synsets in WordNet corresponding to predicate – argument semantic relations

➥ New semantic relations based on predicate – argument structure (resource)

➥ Leseva, Svetlozara, Ivelina Stoyanova, Hristina Kukova, Maria Todorova. Integrating Subcategorisation Information in WordNet. / Светлозара Лесева, Ивелина Стоянова, Христина Кукова, Мария Тодорова. Интегриране на субкатегоризационна информация в релационната структура на Уърднет. – сп. „Български език“, кн. 2, 2018.


Work package 5: Representation of conceptual frames of semantic relations between classes of synsets


Definition of conceptual frames representing the possible combinations between a semantic relation, a verb synset from a particular semantic class and a set of noun synsets from (a) particular semantic class(es)

➥ Hierarchical representation of frame elements and relations. Hierarchy of selectional restrictions and the corresponding sets of noun synsets. (resource)

➥ Assigned semantical frames to 13,226 verb synsets (with 5,025 manually verified). Manually verified entries are labelled by “0++”. (resource)

➥ Leseva, S., Stoyanova, I. Enhancing Conceptual Description through Resource Linking and Exploration of Semantic Relations. In: Proceedings of the 10th Global WordNet Conference, Oficyna Wydawnicza Politechniki Wrocławskiej, 2019, p. 280-289. pdf

➥ Stoyanova, I., Leseva, S. A Structural Approach to Enhancing WordNet with Conceptual Frame Semantics. Proceedings of Recent Advances in Natural Language Processing, Varna, Bulgaria, Sep 2–4 2019, 2019, p. 629-637. pdf

➥ Leseva, S., Stoyanova, I., Todorova, M., Kukova, H. Frame Specialisation Motivated by Inter-Frame Relations in FrameNet. In: Proceedings of the 14th International Conference on Linguistic Resources and Tools for Natural Language Processing, Cluj-Napoca, 18-20 November 2019, Editura Universității „Alexandru Ioan Cuza” din Iași, 2019, 167-178.

➥ Leseva, Svetlozara, Ivelina Stoyanova, Maria Todorova, Hristina Kukova. A Theoretical Overview of Conceptual Frames and Semantic Restrictions on Frame Elements. – В: Балканско езикознание/Linguistique Balkanique, LVIII, 2, 2019, с. 172-186. ISSN 0324-1653. (SCOPUS). pdf pdf

➥ Leseva, Svetlozara, Ivelina Stoyanova, Maria Todorova, Hristina Kukova. A Semantic Description of the Combinability between Verbs and Nouns (on Material from Bulgarian and English). – В: Чуждоезиково обучение, 47, 2, 2020, с. 115 – 128. (Web of Science) резюме

➥ Leseva, S., I. Stoyanova, M. Todorova, H. Kukova. Putting Pieces Together: Predicate-Argument Relations and Selectional Preferences. – In: Koeva, S. (ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations, Professor Marin Drinov Publishing House of BAS, 2020, ISBN 978-619-245-057-1, pp. 49 – 86.


Formulation of conceptual frames representing the possible combinations between a semantic relation, a noun synset from a particular semantic class and a set of noun synsets from (a) particular semantic class(es)

➥ There are 21 new semantic relations defined between noun synsets. 2,814 noun synsets have been connected by experts via 589 relations. (resource)

➥ Dimitrova, T., V. Stefanova. On Hidden Semantic Relations between Nouns in WordNet. In: Proceedings of the Tenth Global WordNet Conference (July 23–27, 2019, Wrocław (Poland), Wroclaw: Oficyna Wydawnicza Politechniki Wrocławskiej, 2019, ISBN 978-83-7493-108-3, 54-63. pdf

➥ Stefanova, Valentina, Tsvetana Dimitrova. About the Participles in the Bulgarian WordNet. / Стефанова, Валентина, Цветана Димитрова. За причастията в Българския Уърднет. – В: Доклади от Международната годишна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 2020 г.). Т. 2, София: Издателство на БАН „Проф. „Марин Дринов“, 2020, с. 224-232. ISSN 2683-118Х (print); ISSN 2683-1198 (online). (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) pdf


Representation of the defined conceptual frames as relations in WordNet, as a result of which WordNet will be enriched with a dense network of semantic relations

➥ Verified and extended models of the Pattern Dictionary of English Verbs – PDEV are added to the XML file of the Princeton WordNet and is available under the Creative Commons license. (resource)

➥ Koeva, S., D. Hristov, T. Dimitrova, V. Stefanova. Enriching Wordnet with Frame Semantics. Доклади от Международната годишна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 14 – 15 май 2019 г.), София: Издателство на БАН „Проф. Марин Дринов“, 2019, ISBN 978-954-322-987-1, 300-308. pdf

➥ Koeva, S., T. Dimitrova, V. Stefanova, D. Hristov. Towards Conceptual Frames. Чуждоезиково обучение, 46, 6, 2019, ISSN 1314–8508 (Online); 0205–1834 (Print), 551-564. pdf

➥ Koeva, S. Semantic Relations and Conceptual Frames (preface). – In: Koeva, S. (ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations, Professor Marin Drinov Publishing House of BAS, 2020, ISBN 978-619-245-057-1, pp. 7–20. pdf


Work package 6: Data consistency checks


Development of validation tests and automated procedures to ensure a consistent representation of the ontology of semantic classes of nouns and verbs in WordNet

➥ Dimitrova, Tsvetana. On WordNet Semantic Classes: Is the Sum Always Bigger? – In: Proceedings of the Fourth International Conference “Computational Linguistics in Bulgaria” (CLIB 2020). Institute for Bulgarian Language – Bulgarian Academy of Sciences, 2020, 177-185. ISSN: 2367-5675. pdf


Development of validation tests and automated procedures to ensure a consistent representation of the obtained dense network of semantic relations in WordNet

➥ Koeva, Svetla, Valentina Stefanova. Meronymy in WordNet: Defining Subrelations. / Коева, Светла, Валентина Стефанова. Меронимията в Уърднет: дефиниране на субрелации. – В: Доклади от Международната годишна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 2020 г.). Т. 2, София: Издателство на БАН „Проф. „Марин Дринов“, 2020, с. 212-223. ISSN 2683-118Х (print); ISSN 2683-1198 (online). (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) pdf


Development of validation tests and automated procedures to ensure a consistent representation of conceptual frames

➥ Leseva, Svetlozara, Ivelina Stoyanova. Consistency Evaluation towards Enhancing the Conceptual Representation of Verbs in WordNet. – In: Proceedings of the Fourth International Conference “Computational Linguistics in Bulgaria” (CLIB 2020), Institute for Bulgarian Language, 2020, pp. 165-175. ISSN 2367-5675 (online). (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) pdf

➥ Leseva, S., I. Stoyanova. Beyond Lexical and Semantic Resources: Linking WordNet with FrameNet and Enhancing Synsets with Conceptual Frames. – In: Koeva, S. (ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations, Professor Marin Drinov Publishing House of BAS, 2020, ISBN 978-619-245-057-1, pp. 7–20. pdf

➥ Leseva, Svetlozara, Ivelina Stoyanova. Beyond Lexical Resources: Validation of Conceptual Description in Corpus Data. / Лесева, Светлозара, Ивелина Стоянова. Отвъд лексикалните ресурси: валидиране на концептуалното описание в корпусни данни. – В: Доклади от Международната годишна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 2020 г.). Т. 2, София: Издателство на БАН „Проф. „Марин Дринов“, 2020, с. 241-249. ISSN 2683-118Х (print); ISSN 2683-1198 (online). (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) pdf



Collection

Koeva, Svetla. (Ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations. Sofia: Professor Marin Drinov Publishing House of BAS, 2020, 121 p. ISBN: 978-619-245-057-1. DOI: 10.7546/TSN.2020 (РИНЦ, CEЕOL) ➥ pdf

Papers and Studies

In SCOPUS

Koeva, Svetla, Tsvetana Dimitrova, Valentina Stefanova, Dimitar Hristov. Mapping WordNet concepts with CPA ontology. – In: Proceedings of the 9th Global WordNet Conference (GWC’2018), Global Wordnet Association, Singapore, 2018, pp. 70-77. ISBN 978-981-11-7087-4. (SCOPUS). ➥ pdf

Stoyanova, Ivelina, Svetlozara Leseva. A Structural Approach to Enhancing WordNet with Conceptual Frame Semantics. – In: Proceedings of Recent Advances in Natural Language Processing, Varna, Bulgaria, Sep 2–4 2019, 2019, pp. 629-637. ISBN 978-954-452-056-4, ISSN 2603-2813. SJR (SCOPUS):0.143. ➥ pdf

Dimitrova, Tsvetana, Valentina Stefanova. On Hidden Semantic Relations between Nouns in WordNet. – In: Proceedings of the Tenth Global WordNet Conference (July 23–27, 2019, Wrocław (Poland). Wroclaw: Oficyna Wydawnicza Politechniki Wrocławskiej, 2019, pp. 54-63. ISBN 978-83-7493-108-3. (SCOPUS) ➥ pdf

Leseva, Svetlozara, Ivelina Stoyanova. Enhancing Conceptual Description through Resource Linking and Exploration of Semantic Relations. – In: Proceedings of the Tenth Global Wordnet Conference (July 23–27, 2019, Wrocław (Poland). Wroclaw: Oficyna Wydawnicza Politechniki Wrocławskiej, 2019, pp. 280-289. ISBN 978-83-7493-108-3. (SCOPUS). pdf ➥ pdf

Leseva, Svetlozara, Ivelina Stoyanova, Maria Todorova, Hristina Kukova. A Theoretical Overview of Conceptual Frames and Semantic Restrictions on Frame Elements. – В: Балканско езикознание/Linguistique Balkanique, LVIII, 2, 2019, с. 172-186. ISSN 0324-1653. (SCOPUS). pdf ➥ pdf

In Web of Science

Piasecki, Maciej, Svetla Koeva. WordNet Relations in the Bulgarian-Polish Bilingual Perspective. – In: Доклади от Международната юбилейна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 15 – 16 май 2017 година). Т. 1, София: Институт за български език „Проф. Любомир Андрейчин“, 2017, с. 285-298. ISBN 978-954-924-899-9. (Web of Science). ➥ pdf

Leseva, Svetlozara, Ivelina Stoyanova, Maria Todorova. Classifying Verbs in WordNet by Harnessing Semantic Resources. – In: Proceedings of the Third International Conference Computational Linguistics in Bulgaria (CLIB 2018). Sofia: The Institute for Bulgarian Language, 2018, pp. 115-125. ISSN 2367-5675, (Web of Science). ➥ pdf

Stoyanova, Ivelina. Factors and Features. Determining the Inheritance of Semantic Primes between Verbs and Nouns within WordNet. – In: Proceedings of the Third International Conference Computational Linguistics in Bulgaria (CLIB 2018). Sofia: The Institute for Bulgarian Language, 2018, pp. 135-145. ISSN 2367-5675. pdf (Web of Science) ➥ pdf

Koeva, Svetla, Valentina Stefanova, Dimitar Hristov. Semantic Relations Within Multiple Hypernymy in WordNet. / Семантични релации в рамките на многократната хиперонимия в УърдНет. – В: сп. „Чуждоезиково обучение“, кн. 4, 2018. (Web of Science). ➥ pdf

Koeva, Svetla, Tsvetana Dimitrova, Valentina Stefanova, Dimitar Hristov. Towards Conceptual Frames. – В: Чуждоезиково обучение, 46, 6, 2019, с. 551-564. ISSN 1314-8508 (online); 0205-1834 (print). (Web of Science). ➥ pdf

Koeva, Svetla. Clauses in the Structure of the Sentence in Bulgarian. / Прости изречения в състава на сложното в български. Релации между присъединяваща част глагол и комплемент. – В: Bednarska, K., Kruk, D., Popov, B., Saprikina, O., Speed, T., Szafraniec, K., Terekhova, S., Tsonev, R., Wysocka, A. (Eds.), 2020, Contributions to the 23rd Annual Scientific Conference of the Association of Slavists (Polyslav). Wiesbaden, 2020, pp. 186-194. (Web of Science)

Leseva, Svetlozara, Ivelina Stoyanova, Maria Todorova, Hristina Kukova. A Semantic Description of the Combinability between Verbs and Nouns (on Material from Bulgarian and English). – В: Чуждоезиково обучение, 47, 2, 2020, с. 115 – 128. (Web of Science) ➥ резюме

In ERIH+

Leseva, Svetlozara. Presenting Knowledge in Semantic Networks. / Представяне на знания чрез семантични мрежи. – В: сп. „Български език“, кн. 2, 2018, с. 59-76. ISSN 0005-4283. (ERIH+) ➥ pdf

Stoyanova, Ivelina, Maria Todorova. Semantic Relations: Theoretical and Practical Aspects. / Семантични релации: теоретични и приложни аспекти. – В: сп. „Български език“, кн. 2, 2018, с. 13-40. ISSN 0005-4283. (ERIH+) ➥ pdf

Dimitrova, Tsvetana. Morphosemantic Relations and Agentive Nouns in the Bulgarian WordNet. / Морфосемантични релации и агентивни съществителни в Българския уърднет. – В: сп. „Български език“, кн. 2, 2018, с. 41-58. ISSN 0005-4283. (ERIH+)➥ pdf

Leseva, Svetlozara, Ivelina Stoyanova, Hristina Kukova, Maria Todorova. Integrating Subcategorisation Information in WordNet. / Интегриране на субкатегоризационна информация в релационната структура на УърдНет. – В: сп. „Български език“, кн. 2, 2018, с. 77-99. ISSN 0005-4283. (ERIH+) ➥ pdf

In CEEOL

Koeva, Svetla. Semantic Relations and Conceptual Frames (preface). – In: Koeva, S. (ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations, Professor Marin Drinov Publishing House of BAS, 2020, pp. 7-20. ISBN 978-619-245-057-1. (РИНЦ, CEEOL) pdf ➥ pdf

Leseva, Svetlozara, Ivelina Stoyanova, Maria Todorova, Hristina Kukova. Putting Pieces Together: Predicate-Argument Relations and Selectional Preferences. – In: Koeva, S. (ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations. Sofia: Professor Marin Drinov Publishing House of BAS, 2020, pp. 49-86. ISBN 978-619-245-057-1. (РИНЦ, CEEOL) pdf ➥ pdf

Leseva, Svetlozara, Ivelina Stoyanova. Beyond Lexical and Semantic Resources: Linking WordNet with FrameNet and Enhancing Synsets with Conceptual Frames. – In: Koeva, S. (ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations. Sofia: Professor Marin Drinov Publishing House of BAS, 2020, pp. 21-48. ISBN 978-619-245-057-1. (РИНЦ, CEЕOL) pdf ➥ pdf

Koeva, Svetla, Tsvetana Dimitrova, Valentina Stefanova, Dimitar Hristov. Towards Conceptual Frames. – In: Koeva, S. (ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations. Sofia: Professor Marin Drinov Publishing House of BAS, 2020, pp. 87-120. ISBN 978-619-245-057-1. (РИНЦ, CEЕOL) pdf ➥ pdf

Submitted for indexing

Koeva, Svetla. Complements in Bulgarian / Комплементите в български. – В: Доклади от Международната годишна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 14 – 15 май 2019 година). Ваня Мичева, Диана Благоева, Сия Колковска, Татяна Александрова, Христина Дейкова (отг. ред.), София: Издателство на БАН „Проф. Марин Дринов“, 2019, с. 57-69. ISBN 978-954-322-987-1. (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) pdf ➥ pdf

Koeva, Svetla, Dimitar Hristov, Tsvetana Dimitrova, Valentina Stefanova. Enriching Wordnet with Frame Semantics. – В: Доклади от Международната годишна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 14 – 15 май 2019 година). Ваня Мичева, Диана Благоева, Сия Колковска, Татяна Александрова, Христина Дейкова (отг. ред.), София: Издателство на БАН „Проф. Марин Дринов“, 2019, с. 300-308. ISBN 978-954-322-987-1. (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) ➥ pdf

Leseva, Svetlozara, Ivelina Stoyanova, Maria Todorova, Hristina Kukova. Frame Specialisation Motivated by Inter-Frame Relations in FrameNet. – In: Proceedings of the 14th International Conference on Linguistic Resources and Tools for Natural Language Processing, Cluj-Napoca, 18-20 November 2019, Editura Universității „Alexandru Ioan Cuza” din Iași, 2019, pp. 167-178. ISSN 1843-911X. (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) ➥ pdf

Dimitrova, Tsvetana. On WordNet Semantic Classes: Is the Sum Always Bigger? – In: Proceedings of the Fourth International Conference “Computational Linguistics in Bulgaria” (CLIB 2020). Sofia: Institute for Bulgarian Language, 2020, pp. 177-185. ISSN 2367-5675. (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) ➥ pdf

Leseva, Svetlozara, Ivelina Stoyanova. Consistency Evaluation towards Enhancing the Conceptual Representation of Verbs in WordNet. – In: Proceedings of the Fourth International Conference “Computational Linguistics in Bulgaria” (CLIB 2020), Institute for Bulgarian Language, 2020, pp. 165-175. ISSN 2367-5675 (online). (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) ➥ pdf

Koeva, Svetla, Valentina Stefanova. Meronymy in WordNet: Defining Subrelations. / Меронимията в Уърднет: дефиниране на субрелации. – В: Доклади от Международната годишна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 2020 г.). Т. 2, София: Издателство на БАН „Проф. „Марин Дринов“, 2020, с. 212-223. ISSN 2683-118Х (print); ISSN 2683-1198 (online). (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) ➥ pdf

Leseva, Svetlozara, Ivelina Stoyanova. Beyond Lexical Resources: Validation of Conceptual Description in Corpus Data. / Отвъд лексикалните ресурси: валидиране на концептуалното описание в корпусни данни. – В: Доклади от Международната годишна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 2020 г.). Т. 2, София: Издателство на БАН „Проф. „Марин Дринов“, 2020, с. 241-249. ISSN 2683-118Х (print); ISSN 2683-1198 (online). (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) ➥ pdf

Stefanova, Valentina, Tsvetana Dimitrova. About the Participles in the Bulgarian WordNet. / За причастията в Българския уърднет. – В: Доклади от Международната годишна конференция на Института за български език „Проф. Любомир Андрейчин“ (София, 2020 г.). Т. 2, София: Издателство на БАН „Проф. „Марин Дринов“, 2020, с. 224-232. ISSN 2683-118Х (print); ISSN 2683-1198 (online). (Сборникът с доклади е подаден за оценка в Web of Science и SCOPUS.) ➥ pdf

Overview

Svetlozara Leseva. Presenting Knowledge in Semantic Networks. (A study, 40 pages, peer-reviewed)

Ivelina Stoyanova and Maria Todorova. Semantic Relations: Theoretical and Practical Aspects.

Tsvetana Dimitrova, Valentina Stefanova. Exploring Existing Semantic Classificarions Applying to Verbs and Nouns.

The work by Maciej Piasecki and Svetla Koeva, WordNet Relations in the Bulgarian-Polish Bilingual Perspective, was presented at the International Conference of the Institute of Bulgarian Language (Sofia, 15 – 16 May 2017), the proceedings of which are indexed in Thomson Reuters Conference Proceedings Citation Index.

Svetla Koeva, Tsvetana Dimitrova, Valentina Stefanova and Dimitar Hristov presented their work Mapping WordNet Concepts with CPA Ontology at the Global WordNet Conference inj Singapore (8 – 12 January 2018).

The paper was published in the Conference proceedings (available here), indexed in Thomson Reuters Book Citation Index, Thomson Reuters Conference Proceedings Citation Index, SCOPUS.

Two papers were presented at the international conference Computational Linguistics in Bulgaria – CLIB 2018:

Svetlozara Leseva presented (joint work withIvelina Stoyanova and Maria Todorova) Classifying Verbs in WordNet by Harnessing Semantic Resources focused on the automatic assignment of semantic clases and subclasses and the hierarchical structure of verbs (as defined in WordNet) by employing features from three resources – WordNet, FrameNet and VerbNet.

Ivelina Stoyanova presented Factors and Features Determining the Inheritance of Semantic Primes between Verbs and Nouns within WordNet, an overview of the ways of inheriting semantic features between derivationally related verbs and nouns.

The papers are published in the Conference proceedings CLIB 2018 indexed in the Thomson Reuters Conference Proceedings Citation Index.

* * * * *

In 2019 two works have been presented at the prestigious Global WordNet Conference (23 – 27 July 2019, Poland):

Dimitrova, T., V. Stefanova. On Hidden Semantic Relations between Nouns in WordNet. Global WordNet Conference (July 23–27, 2019, Wrocław (Poland), Wroclaw: Oficyna Wydawnicza Politechniki Wrocławskiej, 2019.

Leseva, S., Stoyanova, I. Enhancing Conceptual Description through Resource Linking and Exploration of Semantic Relations. Global WordNet Conference(July 23–27, 2019, Wrocław (Poland), Wroclaw: Oficyna Wydawnicza Politechniki Wrocławskiej, 2019.

A joint work of Svetla Koeva, Dimitar Hristov, Tsvetana Dimitrova and Valentina Stefanova entitled Enriching Wordnet with Frame Semantics was presented at the Annual Conference of the Institute for Bulgarian Language (Sofia, 14 – 15 May 2019).

Svetla Koeva presented Complements in Bulgarian at the Annual Conference of the Institute for Bulgarian Language (Sofia, 14 – 15 May 2019).

Svetlozara Leseva presented A Structural Approach to Enhancing WordNet with Conceptual Frame Semantics (joint work with ivelina Stoyanoca) at the Conference Recent Advances in Natural Language Processing RANLP 2019 (Varna, 4 September 2019).

Svetla Koeva presented Clauses in the Structure of the Sentence in Bulgarian at the International Academic Conference POLYSLAV-XXIII (Blagoevgrad, 9 – 11 September 2019).

A joint work by Svetlozara Leseva, Ivelina Stoyanova, Maria Todorova and Hristina Kukova entitled Frame Specialisation Motivated by Inter-Frame Relations in FrameNet was presented at the 14th International Conference on Linguistic Resources and Tools for Natural Language Processing, Cluj-Napoca, 18-20 November 2019, Editura Universității „Alexandru Ioan Cuza” din Iași, 2019.

* * * * *

In 2020 two papers were presented in the Special Session on WordNet and Ontologies at the Fourth International Conference Computational Linguistics in Bulgaria (25 – 26 June 2020):

Tsvetana Dimitrova presented On WordNet Semantic Classes: Is the Sum Always Bigger?. Video recording

Ivelina Stoyanova presented Consistency Evaluation towards Enhancing the Conceptual Representation of Verbs in WordNet (joint work with Svetlozara Leseva)
Video recording

Within the project “Towards a Semantic Network Enriched with a Variety of Semantic Relations”, funded by the National Science Fund under agreement ДН 10/3 from 14.12 2016, there have taken place 7 public seminar focused on current tasks and activities, aiming at exchange of ideas and advice from other scientists in the field of semantic description.

18.05.2017, Institute for Bulgarian Language
Presenter: Maciej Piasecki, Wroclaw Technical University (Poland)
Topic: Latest Developments in plWordNet – a Large Wordnet for Polish. Towards plWordNet 4.0.

24.10.2017, Institute for Bulgarian Language
Presenter: Tsvetana Dimitrova
Topic: Semantical Enrichment of Linguistic Description of Nouns in WordNet.
Joint work by Svetla Koeva, Tsvetana Dimitrova, Valentina Stefanova, Dimitar Hristov

21.11.2017, Institute for Bulgarian Language
Presenter: Борислав Ризов
Topic: Exploring Inheritance of Semantic Primitives between Nouns and Verbs..
Joint work by Tinko Tinchev and Borislav Rizov

28.11.2017, Institute for Bulgarian Language,
Presenter: Svetlozara Leseva
Topic: Verb Semantics – Expanding Classes ans Subclasses within WordNet.
Joint work by Svetlozara Leseva, Ivelina Stoyanova, Maria Todorova

28.11.2017, Institute for Bulgarian Language,
Presenter: Ivelina Stoyanova
Topic: Factors and Features Motivating Inheritance of Semantic Primitives between Verbs and Nouns in BulNet.

20.06.2019, Institute for Bulgarian Language
Presenter: Tsvetana Dimitrova
Topic: “Hidden” Semanric Relations between Nouns in WordNet.
Joint work by Tsvetana Dimitrova and Valentina Stefanova.

10.06.2019, Institute for Bulgarian Language
Presenter: Svetlozara Leseva
Topic: Towards Enhancing the Conceptual Description by Linking Resources and Exploring Semantic Relations.
Joint work by Svetlozara Leseva and Ivelina Stoyanova.

On the 28 and 29 May 2018 in the House of Europe and the Headquarters of the Bulgarian Academy of Sciences in Sofia was held the third edition of the International Scientific Conference “Computational Linguistics in Bulgaria – CLIB 2018”. p>

The special session was organized within the project “Semantic network with a wide range of semantic relations” of the Department of Computational Linguistics at the Institute of Bulgarian Language, funded by the National Science Fund.

The aim of the special session on Wordnet and ontologies was to create a forum for sharing research in the field of lexical-semantic networks and ontologies and the interaction and integration between the two types of knowledge representation in resources with different purposes. The experience and results that were shared offered valuable insights for the future of research in this field and are directly relevant to the implementation of the next stage of the project.

The paper by Natalia Lukashevich and Boris Dobrov (Ontologies for Natural Language Processing: the Case of Russian) presented a group of language resources for Russian, RuThes, based on combining Wordnet with thesauri and formal ontologies, with the data presented in a unified format. The obtained resources are used in the field of computational processing of natural language and information retrieval. One of the real applications of the resource is the semi-automatic generation of RuWordNet.

The work, presented by Ranka Stankovic, Milana Mladenovic, Ivan Obradovic, Marko Vitas and Tsvetana Krstev (Resource-based WordNet Augmentation and Enrichment), demonstrates an approach to enriching Serbian WordNet with the help of Serbian-English resources. The method is based on translating and correcting Princeton WordNet definitions into Serbian and automatically selecting candidates for members of synonymous sets from lists of translation equivalents derived from bilingual resources. An evaluation of the results is presented, taking into account the volume of adjustments made by experts on the automatically created version.

The paper of the colleagues from Romania (Maria Mitrofan, Verjinika Barbu Mititelu, Grigorina Mitrofan – A Pilot Study for Enriching the Romanian WordNet with Medical Terms) presented a pilot study aimed at enriching the Romanian wordnet with specialized vocabulary, in particular medical terminology. The article examines the integration of knowledge from the medical thesaurus SNOMED CT in the hierarchical relational structure of Wordnet and presents problematic cases related to the different organisation of knowledge in the two resources.

The paper on Classifying Verbs in WordNet by Harnessing Semantic Resources (Svetlozara Leseva, Ivelina Stoyanova and Maria Todorova) presented a classification of verbs (as defined in Wordnet), created automatically by combining the advantages of three semantic resources – WordNet itself and its branched hierarchical structure, the rich and granular semantic description and taxonomic relations in FrameNet, and the more generalised semantically and syntactically based description in VerbNet. Based on the alignment between the three resources and their internal structure, a classification is derived, whose categories (semantic classes) are transferred from the frames (conceptual structures) in FrameNet, structured according to the hierarchical relations in Wordnet and FrameNet, reflected in the taxonomy of the two resources.

Ivelina Stoyanova’s report on Factors and Features Determining the Inheritance of Semantic Primes between Verbs and Nouns within WordNet explores the mechanisms for inheriting semantic properties between derivationally connected verbs and nouns and identifies three types of inheritance between the semantic primitives of verbs and nouns: universal – independent from the argumentative structure of the verb, which can be eventful and circumstantial; general – characteristic for whole classes of verbs (e.g., agentive / non-agentive); specific for particular verbs – depending on the argument structure (as presented in resources such as VerbNet and FrameNet). The paper proposes possible expansion of the coverage of semantic relations based on information about the argument structure and discusses the regularities in inheriting semantic characteristics from verbs to nouns and their application for expanding WordNet with semantic sets, for different validations of the data, etc.

During the session, a demonstration of the web-based system for editing and visualization of WordNet, Hydra, was presented (Borislav Rizov and Tsvetana Dimitrova – Online Editor for WordNets). The functionalities of the system allow editing of synonymous sets, including adding / deleting synonyms, compiling and editing the definition, examples and other information, adding or removing relations, etc.

The high scientific value of the works accepted for CLIB 2018 was guaranteed by the selection procedure through double blind review. Each article was evaluated by three independent reviewers, prominent experts in the relevant scientific field. Continuing the tradition of submitting the Proceedings for indexing in prestigious databases of scientific publications, currently they are indeced in Web of Science.

The media partner of the Conference, the National Publishing House for Education and Science Az Buki, played a big role in covering the event, and also published an extensive interview with Prof. Dr. Ruslan Mitkov in Az Buki newspaper (in Bulgarian).

Conference Proceedings, programme and photos are available online on the CLIB 2018 webpage.

Sponsors of CLIB 2018 are the companies A Data Pro, Identrics and Documaster; media partner is the National Publishing House for Education and Science Az Buki.



The fourth edition of the International Conference on Computational Linguistics in Bulgaria (CLIB 2020: https://dcl.bas.bg/clib/) took place on 25 and 26 June 2020. The forum was held in a mixed format – in person (at the Prof. Marin Drinov Hall of the Bulgarian Academy of Sciences) and online by means of a video conferencing platform. The conference was broadcast online on the YouTube channel of the Department of Computational Linguistics.

The mission of the conference, which was first held in 2014, is to enhance and expand the cooperation between Bulgarian researchers working in the field of computational linguistics in Bulgaria and around the world and foreign scientists who develop language technologies applicable to Bulgarian.
85 participants from more than 10 countries – Bulgaria, Great Britain, Germany, Qatar, Romania, Russia, Serbia, Slovakia, Turkey, Ukraine, Switzerland, Japan – registered to CLIB 2020. The interest in the conference has been growing even after its end as indicated by the number of times the conference videos were watched – close to 500 times until now.

Three plenary talks were delivered on the two days of the conference by leading Bulgarian scientists, who presented their achievements in topical areas of computational linguistics: sense disambiguation in image collections (Prof. Dr. Sc. Galia Angelova), natural language processing of specialised (medical) documents with a focus on information retrieval (Assoc. Prof. Svetla Boytcheva) and advanced fake news detection (Dr. Preslav Nakov). A special demonstration of the European Language Grid was presented by the project’s coordinator Dr. Georg Rehm.

The main conference included 14 talks presenting the work of researchers from Bulgaria, Romania, Russia, Serbia, Turkey, Japan, Switzerland. The second edition of the Special Session on Wordnets and Ontologies took place on the second day of the conference, featuring 5 talks.

The conference is organised by the Department of Computational Linguistics at the Institute for Bulgarian Language Prof. Lyubomir Andreychin and the Institute of Information and Communication Technologies, both at the Bulgarian Academy of Sciences.

CLIB 2020 was supported by the European Language Resources Association (ELRA) and the technological company Mozaika. The Special Session on Wordnets and Ontologies was supported within the project A Semantic Network Enriched with a Variety of Semantic Relations funded by the National Science Fund.

The collection comprises of three studies presenting project results – the integration of knowledge from various semantic resources for the purpose of enriching the semantic description of verbs within WordNet.

➥ Koeva, S. (Ed.) Towards a Semantic Network Enriched with a Variety of Semantic Relations. Prof. Marin Drinov Academic Publishing House of Bulgarian Academy of Sciences. Sofia, 2020. ISBN 978-619-245-057-1. DOI: 10.7546/TSN.2020.

Copyright © 2015 Department of computational linguistics. All rights reserved.