Papers by Nikola Ljubešić
The GINCO Training Dataset for Web Genre Identification of Documents Out in the Wild (2022.lrec-1)
Copied to clipboard
| Challenge: | GINCO is a new training dataset for automatic genre identification based on 1,125 crawled Slovenian web documents that consist of 650,000 words. |
| Approach: | They propose to use 1,125 crawled Slovenian web documents to train a new genre classification system based on a GINCO training dataset . |
| Outcome: | The proposed classifiers perform better on the 1,125 crawled Slovenian web documents than the existing models and achieve higher scores on the task. |
A Lightweight Approach to a Giga-Corpus of Historical Periodicals: The Story of a Slovenian Historical Newspaper Collection (2024.lrec-main)
Copied to clipboard
| Challenge: | a curated corpus of Slovenian historical newspapers is a complex undertaking requiring multiple steps to prepare . a shoestring budget is required to produce a corpus that is billion-words in size . |
| Approach: | They propose a lightweight approach to producing high-quality corpora using OCR . they use noisy OCR-ed data from the National and University Library of Slovenia . |
| Outcome: | The proposed method produces a billion-word giga-corpus of Slovenian historical newspapers from the 18th, 19th and 20th centuries on a shoestring budget. |
Universal NER: A Gold-Standard Multilingual Named Entity Recognition Benchmark (2024.naacl-long)
Copied to clipboard
Stephen Mayhew, Terra Blevins, Shuheng Liu, Marek Suppa, Hila Gonen, Joseph Marvin Imperial, Börje Karlsson, Peiqin Lin, Nikola Ljubešić, Lester James Miranda, Barbara Plank, Arij Riabi, Yuval Pinter
| Challenge: | In named entity recognition, the majority of annotation efforts are centered on English, and cross-lingual transfer performance remains brittle. |
| Approach: | They propose to develop gold-standard named entity recognition benchmarks in many languages using a cross-lingual consistent schema. |
| Outcome: | The proposed benchmarks will be released to the public in 2022 . they will provide baselines on in-language and cross-lingual learning settings. |
Gigafida 2.0: The Reference Corpus of Written Standard Slovene (2020.lrec-1)
Copied to clipboard
Simon Krek, Špela Arhar Holdt, Tomaž Erjavec, Jaka Čibej, Andraz Repar, Polona Gantar, Nikola Ljubešić, Iztok Kosem, Kaja Dobrovoljc
| Challenge: | Gigafida reference corpus of Slovene is updated with new material and tools . focus of upgrade was on transformation from general reference corp to standard reference corp . |
| Approach: | We present a new version of the Gigafida reference corpus of Slovene . the upgrade includes new material and better tools for annotating it . |
| Outcome: | The new version of the Gigafida reference corpus of Slovene is described . the whole Gigido corpus was deduplicated for the first time . |
The ParlaSent Multilingual Training Dataset for Sentiment Identification in Parliamentary Proceedings (2024.lrec-main)
Copied to clipboard
| Challenge: | The paper presents a new training dataset of sentences in 7 languages, manually annotated for sentiment, which is used in a series of experiments focused on training a robust sentiment identifier for parliamentary proceedings. |
| Approach: | They propose to use a dataset of sentences manually annotated for sentiment to train a robust sentiment identifier for parliamentary proceedings. |
| Outcome: | The proposed model performs very well on languages not seen during fine-tuning and additional fine- tuning data from other languages significantly improves the target parliament’s results. |
Bleaching Text: Abstract Features for Cross-lingual Gender Prediction (P18-2)
Copied to clipboard
| Challenge: | Existing gender prediction models rely on lexical and social network features to capture style beyond topic. |
| Approach: | They propose an alternative to lexical bleaching, i.e., transforming lexicals into more abstract features. |
| Outcome: | The proposed model performs similar to lexical models, but is less language-, topic-, and platform dependent. |
CLASSLA-web: Comparable Web Corpora of South Slavic Languages Enriched with Linguistic and Genre Annotation (2024.lrec-main)
Copied to clipboard
| Challenge: | Using a similar crawling setup, the corpora are comparable across the entire South Slavic language space. |
| Approach: | They propose to collect 13 billion tokens of texts from 26 million documents . they are linguistically annotated with a CLASSLA-Stanza pipeline and enriched with document-level genre information via a Transformer-based multilingual classifier. |
| Outcome: | The corpora are linguistically annotated with the state-of-the-art CLASSLA-Stanza linguistic processing pipeline and enriched with document-level genre information via the Transformer-based multilingual X-GENRE classifier. |
Gos 2: A New Reference Corpus of Spoken Slovenian (2024.lrec-main)
Copied to clipboard
| Challenge: | a new corpus of spoken Slovenian has been added to the Gos reference corpus . the corpus is now more than double the original size of 300 hours, 2.4 million words . |
| Approach: | They propose to add speech recordings and transcriptions from two related initiatives, the Gos VideoLectures corpus of public academic speech, and the Artur speech recognition database. |
| Outcome: | The new corpus is double the original size and contains 2.4 million words . it includes speech recordings and transcriptions from two related initiatives . |
CoSimLex: A Resource for Evaluating Graded Word Similarity in Context (2020.lrec-1)
Copied to clipboard
Carlos Santos Armendariz, Matthew Purver, Matej Ulčar, Senja Pollak, Nikola Ljubešić, Mark Granroth-Wilding
| Challenge: | Existing methods to evaluate word embeddings ignore context and treat words in isolation. |
| Approach: | They propose to build a new word embeddings-based dataset that provides context-dependent similarity measures. |
| Outcome: | The proposed dataset provides context-dependent similarity measures and covers a well-resourced language (English) but a number of less-resource languages. |
Do Language Models Care about Text Quality? Evaluating Web-Crawled Corpora across 11 Languages (2024.lrec-main)
Copied to clipboard
Rik van Noord, Taja Kuzman, Peter Rupnik, Nikola Ljubešić, Miquel Esplà-Gomis, Gema Ramírez-Sánchez, Antonio Toral
| Challenge: | Large, curated, web-crawled corpora play a vital role in training language models . however, relatively little attention has been given to the quality of these corporata . |
| Approach: | They compare four of the currently most relevant large, web-crawled corpora across eleven lower-resourced European languages to evaluate their quality. |
| Outcome: | The CC100 corpus achieves the highest scores on the tests in 11 lower-resourced European languages. |