• Digital Humanities and Cultural Heritage
• 2014-2023close
• Open Access close
• Other research products close
• Croatian close
• Digital Humanities and Cultural Her... close

The script for the course Sculpture 2 at the Graduate Study of the Department of Sculpture of the Art Academy in Split continues on the script for the course Sculpture 1 from the previous semester. Fifteen teaching units are arranged in sequences consistent with real sculptural practice. The possibility of perception and articulation of new content is one of the most important aspects of the development of the author's work process. The emphasis is on its openness and improvement, as an upgrade of knowledge from the previous semester. Throughout the script, the basic structure of the creative process is determined, but now at a more demanding level. The attention paid to the language of art, as a fundamental and unifying aspect of the work, this semester is further emphasized and presented by a series of selected visual examples, as well as examples of quotations from authors that can be interpreted as the basis of the artist’s statement. The script makes it easier for students to achieve the expected learning outcomes, i.e. awareness of the full process of realization of contemporary author's sculpture, from conceiving an idea to performing the work, and thinking about its layout in exhibition space. Skripta za kolegij Kiparstvo 2 na Diplomskom studiju Odsjeka kiparstvo Umjetničke akademije u Splitu nastavlja se na skriptu za kolegij iz prethodnog semestra, Kiparstvo 1. Petnaest nastavnih jedinica nižu se slijedom sukladnim realnoj kiparskoj praksi, što predstavlja važno iskustvo za studente, osobito u kolegiju Kiparstvo 2 koji nastavno na prethodni kolegij omogućuje studentima daljnje razvijanje znanja i iskustva. Mogućnost percepcije i artikulacije novih sadržaja je jedan od najvažnijih aspekata razvoja procesa autorskog rada. Naglasak je na njegovoj otvorenosti i unaprjeđivanju, kao nadogradnji znanja iz prethodnog semestra. Kroz čitavu skriptu se utvrđuje osnovna struktura stvaralačkog procesa, no sada na zahtjevnijoj razini. Pažnja posvećena likovnom jeziku, kao temeljnom i objedinjujućem aspektu rada, u ovom semestru je dodatno naglašena i prezentirana nizom odabranih vizualnih primjera, kao i primjera citata autora koji se mogu tumačiti kao temelji izjave o radu. Skripta olakšava studentima ostvarivanje očekivanih ishoda učenja, odnosno osvješćivanje potpunog procesa realizacije suvremene autorske skulpture, od osmišljavanja ideje do izvedbe rada i promišljanja njegova postava u prostoru.

The lexicon contains manual translations of the NRC Emotion Lexicon (http://saifmohammad.com/WebPages/NRC-Emotion-Lexicon.htm) that encodes the sentiment of a word (positive, negative) and its emotion association (anger, anticipation, disgust, fear, joy, sadness, surprise, trust) for Croatian, Dutch and Slovene with a binary schema. Manual translations were produced by inspecting and correcting the automatic translations from English provided with the original lexicon. While translations to all 14,182 entries are provided for Slovene and Croatian, only translations for the 6,468 entries that have any sentiment or emotion associated with the word are given for Dutch. For English entries, please refer to the original NRC Emotion Lexicon.

Trilingual BERT (Bidirectional Encoder Representations from Transformers) model, trained on Croatian, Slovenian, and English data. State of the art tool representing words/tokens as contextually dependent word embeddings, used for various NLP classification tasks by finetuning the model end-to-end. CroSloEngual BERT are neural network weights and configuration files in pytorch format (ie. to be used with pytorch library).

hrLex is a large inflectional lexicon of Croatian language where each entry consists of a (wordform, lemma, MSD, MSD features, UPOS, morphological features, frequency, per-million frequency) 8-tuple. The (wordform, lemma, MSD) triple frequencies are calculated on the hrWaC v2.2 corpus. The MSD tagset follows the MULTEXT-East V6 tagset for the Serbo-Croatian macro-language available at http://nl.ijs.si/ME/V6/msd/html/msd-hbs.html. The UPOS + morphological features follow the UD v2 specifications available at http://universaldependencies.org/guidelines.html.

Trilingual BERT (Bidirectional Encoder Representations from Transformers) model, trained on Croatian, Slovenian, and English data. State of the art tool representing words/tokens as contextually dependent word embeddings, used for various NLP classification tasks by finetuning the model end-to-end. CroSloEngual BERT are neural network weights and configuration files in pytorch format (i.e. to be used with pytorch library). Changes in version 1.1: fixed vocab.txt file, as previous verson had an error causing very bad results during fine-tuning and/or evaluation.

The hr500k training corpus contains about 500,000 tokens manually annotated on the levels of tokenisation, sentence segmentation, morphosyntactic tagging, lemmatisation and named entities. About half of the corpus is also manually annotated with syntactic dependencies. A subset of the syntactically annotated corpus is also annotated for multi-word expressions. Furthermore, about a fifth of the corpus is annotated with semantic role labels. The annotation formalisms followed in the hr500k corpus are (1) the MULTEXT-East V6 morphosyntactic specifications for the Serbo-Croatian macro-language, https://nl.ijs.si/ME/V6/msd/, (2) the UDv2 Guidelines, http://universaldependencies.org/guidelines.html, (3) the Janes annotation guidelines for named entities, https://nl.ijs.si/janes/wp-content/uploads/2017/09/SlovenianNER-eng-v1.1.pdf, (4) the PARSEME guidelines for annotating multi-word expressions, https://parsemefr.lis-lab.fr/parseme-st-guidelines/1.3/ and (4) the semantic role labelling annotation protocol for Slovenian and Croatian, https://www.sdjt.si/wp/wp-content/uploads/2018/09/JTDH-2018_Gantar-et-al_Towards-Semantic-Role-Labeling-in-Slovene-and-Croatian.pdf. Different to the previous version of the dataset, it is now encoded in the conllup format, as are other linguistic training datasets for Croatian and Serbian. The PARSEME multi-word expression annotation layer was added as well, together with countless corrections of labels on all available levels. The continuous improvement of this dataset is led by the CLASSLA knowledge centre for South Slavic languages (https://www.clarin.si/info/k-centre/) and the ReLDI Centre Belgrade (https://reldi.spur.uzh.ch).

The model for lemmatisation of non-standard Croatian was built with the CLASSLA-StanfordNLP tool (https://github.com/clarinsi/classla-stanfordnlp) by training on the hr500k training corpus (http://hdl.handle.net/11356/1183), the ReLDI-NormTagNER-hr corpus (http://hdl.handle.net/11356/1241), the RAPUT corpus (https://www.aclweb.org/anthology/L16-1513/) and the ReLDI-NormTagNER-sr corpus (http://hdl.handle.net/11356/1240), using the hrLex inflectional lexicon (http://hdl.handle.net/11356/1232). These corpora were additionally augmented for handling missing diacritics by repeating parts of the corpora with diacritics removed. The estimated F1 of the lemma annotations is ~97.54. The difference to the previous version of the lemmatizer is that now it relies solely on XPOS annotations, and not on a combination of UPOS, FEATS (lexicon lookup) and XPOS (lemma prediction) annotations.

A comprehensive corpus of news articles on the topic of language, published in major Croatian daily newspapers and news portals in the five-year period of January 1, 2015 - January 1, 2020. The corpus is designed to facilitate research on metalanguage (‘language about language’), linguistic ideologies, language policy and planning, as well as the specific contemporary debates on language defining, naming, and standardisation, ongoing in post-Yugoslav societies. The corpus has been tagged using the CLASSLA-StanfordNLP models for morphosyntactic annotation and lemmatisation of standard Croatian. The corpus is available in plain text version, XML with full metadata, and tagged CONLL-U format. MetaLangNEWS-Hr is complemented with a separate corpus of citizen metalanguage comments, i.e. online comments to the news articles, available as MetaLangNEWS-COMMENTS-Hr (http://hdl.handle.net/11356/1370). Parallel versions from Slovenia (http://hdl.handle.net/11356/1360) and Serbia (http://hdl.handle.net/11356/1371) are also available.

The dataset of user comments provided for research purposes for the EMBEDDIA, a Horizon 2020 project, extracted from the database of user comments from the 24sata.hr news portal. The 24sata.hr is the largest-circulation daily newspaper in Croatia, reaching on average 2 million readers daily. The dataset provides the comments metadata including the link to the relevant article, the ID of the comment author (anonymized), and timestamp. The comments are also labelled if they are blocked by human moderators. Description of the Datasets. The 24sata dataset consists of 11 columns and 21548192 rows. Each row represents one user comment on the 24sata news portal. Comments are added by registered users below the published news article. Columns: 'comment_id' - The internal id of the comment. Unique for each row. 'user_id' - The internal id of the user writing the comment. Unique for each user. '0' for all blocked comments. 'content' - The content (text) of the user comment. 'site' - The site the comment came from. 'reply_to_id' - The 'comment_id' of the parent comment - if this comment was intended as a reply. 'created_date' - The date the comment was created. 'last_change' - The date the comment was last edited. 'article_id' - A public id of the article where this comment was posted. The article itself can be accessed by appending article_id to the site. So an article with article_id 614684 and site 'www.24sata.hr' can be found on 'www.24sata.hr/a-614684'. (note the added 'a-' before the article name) 'infringed_on_rule' - If the user has infringed on rules with this comment, id of the rule is given. The description of the rules is given below. 'like_counts' - A number of times other users have voted in favour of this comment, similar to the Like button. 'dislike_counts' - A number of times other users have voted against this comment, opposite of the Like button.

The hr500k training corpus contains about 500,000 tokens manually annotated on the levels of tokenisation, sentence segmentation, morphosyntactic tagging, lemmatisation and named entities. About half of the corpus is also manually annotated with syntactic dependencies. Furthermore, about a fifth of the corpus is annotated with semantic role labels. The annotations (and other aspects) of the hr500k corpus are documented in the teiHeader and back element of the TEI encoded corpus. In short, they follow (1) the MULTEXT-East V5 morphosyntactic specifications for Croatian, https://nl.ijs.si/ME/V5/msd/, (2) the UDv2 Guidelines, http://universaldependencies.org/guidelines.html, and (3) the Janes annotation guidelines for named entities, https://nl.ijs.si/janes/wp-content/uploads/2017/09/SlovenianNER-eng-v1.1.pdf, while (4) the semantic role labelling annotation guidelines are currently in the publication process.