Workshop on Ukrainian Natural Language Processing (2024)


up

pdf (full)
bib (full)
Proceedings of the Third Ukrainian Natural Language Processing Workshop (UNLP) @ LREC-COLING 2024

pdf bib
Proceedings of the Third Ukrainian Natural Language Processing Workshop (UNLP) @ LREC-COLING 2024
Mariana Romanyshyn | Nataliia Romanyshyn | Andrii Hlybovets | Oleksii Ignatenko

pdf bib
A Contemporary News Corpus of Ukrainian (CNC-UA): Compilation, Annotation, Publication
Stefan Fischer | Kateryna Haidarzhyi | Jörg Knappen | Olha Polishchuk | Yuliya Stodolinska | Elke Teich

We present a corpus of contemporary Ukrainian news articles published between 2019 and 2022 on the news website of the national public broadcaster of Ukraine, commonly known as SUSPILNE. The current release comprises 87 210 364 words in 292 955 texts. Texts are annotated with titles and their time of publication. In addition, the corpus has been linguistically annotated at the token level with a dependency parser. To provide further aspects for investigation, a topic model was trained on the corpus. The corpus is hosted (Fischer et al., 2023) at the Saarbrücken CLARIN center under a CC BY-NC-ND 4.0 license and available in two tab-separated formats: CoNLL-U (de Marneffe et al., 2021) and vertical text format (VRT) as used by the IMS Open Corpus Workbench (CWB; Evert and Hardie, 2011) and CQPweb (Hardie, 2012). We show examples of using the CQPweb interface, which allows to extract the quantitative data necessary for distributional and collocation analyses of the CNC-UA. As the CNC-UA contains news texts documenting recent events, it is highly relevant not only for linguistic analyses of the modern Ukrainian language but also for socio-cultural and political studies.

pdf bib
Introducing the Djinni Recruitment Dataset: A Corpus of Anonymized CVs and Job Postings
Nazarii Drushchak | Mariana Romanyshyn

This paper introduces the Djinni Recruitment Dataset, a large-scale open-source corpus of candidate profiles and job descriptions. With over 150,000 jobs and 230,000 candidates, the dataset includes samples in English and Ukrainian, thereby facilitating advancements in the recruitment domain of natural language processing (NLP) for both languages. It is one of the first open-source corpora in the recruitment domain, opening up new opportunities for AI-driven recruitment technologies and related fields. Notably, the dataset is accessible under the MIT license, encouraging widespread adoption for both scientific research and commercial projects.

pdf bib
Creating Parallel Corpora for Ukrainian: A German-Ukrainian Parallel Corpus (ParaRook||DE-UK)
Maria Shvedova | Arsenii Lukashevskyi

Parallel corpora are currently a popular and vibrantly developing category of linguistic resources, used both in literature and translation studies, as well as in the field of NLP. For Ukrainian, though, there are still not enough significant parallel corpora compiled within a single roof project and made available to the research community. In this paper we present a newly developed resource, the German-Ukrainian Parallel Corpus — ParaRook||DE-UK, searchable online. We describe various issues related to its compilation, text selection, and annotation. The paper also features several examples of how the corpus can be used in linguistic research and translation studies. Using the experience of the German-Ukrainian parallel corpus, parallel corpora for other languages with Ukrainian can be developed.

pdf bib
Introducing NER-UK 2.0: A Rich Corpus of Named Entities for Ukrainian
Dmytro Chaplynskyi | Mariana Romanyshyn

This paper presents NER-UK 2.0, a corpus of texts in the Ukrainian language manually annotated for the named entity recognition task. The corpus contains 560 texts of multiple genres, boasting 21,993 entities in total. The annotation scheme covers 13 entity types, namely location, person name, organization, artifact, document, job title, date, time, period, money, percentage, quantity, and miscellaneous. Such a rich set of entities makes the corpus valuable for training named-entity recognition models in various domains, including news, social media posts, legal documents, and procurement contracts. The paper presents an updated baseline solution for named entity recognition in Ukrainian with 0.89 F1. The corpus is the largest of its kind for the Ukrainian language and is available for download.

pdf bib
Instant Messaging Platforms News Multi-Task Classification for Stance, Sentiment, and Discrimination Detection
Taras Ustyianovych | Denilson Barbosa

In the digital age, geopolitical events frequently catalyze discussions among global web users. Platforms such as social networks and messaging applications serve as vital means for information spreading and acquisition. The Russian aggression against Ukraine has notably intensified online discourse on the matter, drawing a significant audience eager for real-time updates. This surge in online activity inevitably results in the proliferation of content, some of which may be unreliable or manipulative. Given this context, the identification of such content with information distortion is imperative to mitigate bias and promote fairness. However, this task presents considerable challenges, primarily due to the lack of sophisticated language models capable of understanding the nuances and context of texts in low-resource languages, and the scarcity of well-annotated datasets for training such models. To address these gaps, we introduce the TRWU dataset - a meticulously annotated collection of Telegram news about the Russian war in Ukraine gathered starting from January 1, 2022. This paper outlines our methodology for semantic analysis and classification of these messages, aiming to ascertain their bias. Such an approach enhances our ability to detect manipulative and destructive content. Through descriptive statistical analysis, we explore deviations in message sentiment, stance, and metadata across different types of channels and levels of content creation activity. Our findings indicate a predominance of negative sentiment within the dataset. Additionally, our research elucidates distinct differences in the linguistic choices and phraseology among channels, based on their stance towards the war. This study contributes to the broader effort of understanding the spread and mitigating the impact of biased and manipulative content in digital communications.

pdf bib
Setting up the Data Printer with Improved English to Ukrainian Machine Translation
Yurii Paniv | Dmytro Chaplynskyi | Nikita Trynus | Volodymyr Kyrylov

To build large language models for Ukrainian we need to expand our corpora with large amounts of new algorithmic tasks expressed in natural language. Examples of task performance expressed in English are abundant, so with a high-quality translation system our community will be enabled to curate datasets faster. To aid this goal, we introduce a recipe to build a translation system using supervised finetuning of a large pretrained language model with a noisy parallel dataset of 3M pairs of Ukrainian and English sentences followed by a second phase of training using 17K examples selected by k-fold perplexity filtering on another dataset of higher quality. Our decoder-only model named Dragoman beats performance of previous state of the art encoder-decoder models on the FLORES devtest set.

pdf bib
Automated Extraction of Hypo-Hypernym Relations for the Ukrainian WordNet
Nataliia Romanyshyn | Dmytro Chaplynskyi | Mariana Romanyshyn

WordNet is a crucial resource in linguistics and natural language processing, providing a detailed and expansive set of lexico-semantic relationships among words in a language. The trend toward automated construction and expansion of WordNets has become increasingly popular due to the high costs of manual development. This study aims to automate the development of the Ukrainian WordNet, explicitly concentrating on hypo-hypernym relations that are crucial building blocks of the hierarchical structure of WordNet. Utilizing the linking between Princeton WordNet, Wikidata, and multilingual resources from Wikipedia, the proposed approach successfully mapped 17% of Princeton WordNet (PWN) content to Ukrainian Wikipedia. Furthermore, the study introduces three innovative strategies for generating new entries to fill in the gaps of the Ukrainian WordNet: machine translation, the Hypernym Discovery model, and the Hypernym Instruction-Following LLaMA model. The latter model shows a high level of effectiveness, evidenced by a 41.61% performance on the Mean Overlap Coefficient (MOC) metric. With the proposed approach that combines automated techniques with expert human input, we provide a reliable basis for creating the Ukrainian WordNet.

pdf bib
Ukrainian Visual Word Sense Disambiguation Benchmark
Yurii Laba | Yaryna Mohytych | Ivanna Rohulia | Halyna Kyryleyza | Hanna Dydyk-Meush | Oles Dobosevych | Rostyslav Hryniv

This study presents a benchmark for evaluating the Visual Word Sense Disambiguation (Visual-WSD) task in Ukrainian. The main goal of the Visual-WSD task is to identify, with minimal contextual information, the most appropriate representation of a given ambiguous word from a set of ten images. To construct this benchmark, we followed a methodology similar to that proposed by (CITATION), who previously introduced benchmarks for the Visual-WSD task in English, Italian, and Farsi. This approach allows us to incorporate the Ukrainian benchmark into a broader framework for cross-language model performance comparisons. We collected the benchmark data semi-automatically and refined it with input from domain experts. We then assessed eight multilingual and multimodal large language models using this benchmark. All tested models performed worse than the zero-shot CLIP-based baseline model (CITATION) used by (CITATION) for the English Visual-WSD task. Our analysis revealed a significant performance gap in the Visual-WSD task between Ukrainian and English.

pdf bib
The UNLP 2024 Shared Task on Fine-Tuning Large Language Models for Ukrainian
Mariana Romanyshyn | Oleksiy Syvokon | Roman Kyslyi

This paper presents the results of the UNLP 2024 shared task, the first Shared Task on Fine-Tuning Large Language Models for the Ukrainian language. The goal of the task was to facilitate the creation of models that have knowledge of the Ukrainian language, history, and culture, as well as common knowledge, and are capable of generating fluent and accurate responses in Ukrainian. The participants were required to use models with open weights and reasonable size to ensure the reproducibility of the solutions. The participating systems were evaluated using multiple-choice exam questions and manually crafted open questions. Three teams submitted their solutions before the deadline, and two teams submitted papers that were accepted to appear in the UNLP workshop proceedings and are referred to in this report. The Codabench leaderboard is left open for further submissions.

pdf bib
Fine-Tuning and Retrieval Augmented Generation for Question Answering Using Affordable Large Language Models
Tiberiu Boros | Radu Chivereanu | Stefan Dumitrescu | Octavian Purcaru

We present our proposed system named Sherlock to UNLP 2024 Shared Task on Question Answering winning first place. We employ a mix of methods, from using automatically translated datasets to perform supervised fine-tuning and direct preference optimization on instruction-tuned models, to model weight merging and retrieval augmented generation. We present and motivate our chosen sequence of steps, as well as an ablation study to understand the effect of each additional step. The resulting model and code are made publicly available (download links provided in the paper).

pdf bib
From Bytes to Borsch: Fine-Tuning Gemma and Mistral for the Ukrainian Language Representation
Artur Kiulian | Anton Polishko | Mykola Khandoga | Oryna Chubych | Jack Connor | Raghav Ravishankar | Adarsh Shirawalmath

In the rapidly advancing field of AI and NLP, generative large language models (LLMs) stand at the forefront of innovation, showcasing unparalleled abilities in text understanding and generation. However, the limited representation of low-resource languages like Ukrainian poses a notable challenge, restricting the reach and relevance of this technology. Our paper addresses this by fine-tuning the open-source Gemma and Mistral LLMs with Ukrainian datasets, aiming to improve their linguistic proficiency and benchmarking them against other existing models capable of processing Ukrainian language. This endeavor not only aims to mitigate language bias in technology but also promotes inclusivity in the digital realm. Our transparent and reproducible approach encourages further NLP research and development. Additionally, we present the Ukrainian Knowledge and Instruction Dataset (UKID) to aid future efforts in language model fine-tuning. Our research not only advances the field of NLP but also highlights the importance of linguistic diversity in AI, which is crucial for cultural preservation, education, and expanding AI’s global utility. Ultimately, we advocate for a future where technology is inclusive, enabling AI to communicate effectively across all languages, especially those currently underrepresented.

pdf bib
Spivavtor: An Instruction Tuned Ukrainian Text Editing Model
Aman Saini | Artem Chernodub | Vipul Raheja | Vivek Kulkarni

We introduce Spivavtor, a dataset, and instruction-tuned models for text editing focused on the Ukrainian language. Spivavtor is the Ukrainian-focused adaptation of the English-only CoEdIT (Raheja et al., 2023) model. Similar to CoEdIT, Spivavtor performs text editing tasks by following instructions in Ukrainian like “Виправте граматику в цьому реченнi” and “Спростiть це речення” which translate to “Correct the grammar in this sentence” and “Simplify this sentence” in English, respectively. This paper describes the details of the Spivavtor-Instruct dataset and Spivavtor models. We evaluate Spivavtor on a variety of text editing tasks in Ukrainian, such as Grammatical Error Correction (GEC), Text Simplification, Coherence, and Paraphrasing, and demonstrate its superior performance on all of them. We publicly release our best performing models and data as resources to the community to advance further research in this space.

pdf bib
Eval-UA-tion 1.0: Benchmark for Evaluating Ukrainian (Large) Language Models
Serhii Hamotskyi | Anna-Izabella Levbarg | Christian Hänig

In this paper, we introduce Eval-UA-tion, a set of novel Ukrainian-language datasets aimed at evaluating the performance of language models on the Ukrainian language. The tasks include UA-CBT (inspired by the Children’s Book Test, a fill-in-the-gaps type task aimed at gauging the extent to which a story narrative is understood), UP-Titles (where the online newspaper Ukrainska Pravda‘s articles have to be matched to the correct title among 10 similar ones), and LMentry-static-UA/LMES (inspired by the LMentry benchmark, a set of tasks simple to solve for humans but hard for LMs, such as ‘which of these words is longer’ and ‘what is the fifth word of this sentence’). With the exception of UP-Titles, the tasks are built in a way to minimize contamination and use material unlikely to be present in the training sets of language models, and include a split for few-shot model prompting use that minimizes contamination. For each task human and random baselines are provided.

pdf bib
LiBERTa: Advancing Ukrainian Language Modeling through Pre-training from Scratch
Mykola Haltiuk | Aleksander Smywiński-Pohl

Recent advancements in Natural Language Processing (NLP) have spurred remarkable progress in language modeling, predominantly benefiting English. While Ukrainian NLP has long grappled with significant challenges due to limited data and computational resources, recent years have seen a shift with the emergence of new corpora, marking a pivotal moment in addressing these obstacles. This paper introduces LiBERTa Large, the inaugural BERT Large model pre-trained entirely from scratch only on Ukrainian texts. Leveraging extensive multilingual text corpora, including a substantial Ukrainian subset, LiBERTa Large establishes a foundational resource for Ukrainian NLU tasks. Our model outperforms existing multilingual and monolingual models pre-trained from scratch for Ukrainian, demonstrating competitive performance against those relying on cross-lingual transfer from English. This achievement underscores our ability to achieve superior performance through pre-training from scratch with additional enhancements, obviating the need to rely on decisions made for English models to efficiently transfer weights. We establish LiBERTa Large as a robust baseline, paving the way for future advancements in Ukrainian language modeling.

pdf bib
Entity Embellishment Mitigation in LLMs Output with Noisy Synthetic Dataset for Alignment
Svitlana Galeshchuk

The present work focuses on the entity embellishments when named entities are accompanied by additional information that is not supported by the context or the source material. Our paper contributes into mitigating this problem in large language model’s generated texts, summaries in particular, by proposing the approach with synthetic noise injection in the generated samples that are further used for alignment of finetuned LLM. We also challenge the issue of solutions scarcity for low-resourced languages and test our approach with corpora in Ukrainian.

pdf bib
Language-Specific Pruning for Efficient Reduction of Large Language Models
Maksym Shamrai

Delving into pruning techniques is essential to boost the efficiency of Large Language Models (LLMs) by reducing their size and computational demands, resulting in faster and more cost-effective inference. In this work, our key contribution lies in recognizing that LLMs trained on diverse languages manifest distinct language-specific weight distributions. Exploiting this insight, we illustrate that pruning LLMs using language-specific data results in a more potent model compression. Empirical evidence underscores the critical nature of pruning on language-specific data, highlighting a noteworthy impact on the perplexity of Ukrainian texts compared to pruning on English data. The proposed methodology significantly reduces the size of LLaMA, LLaMA 2 and Mistral models while preserving competitive performance. This research underscores the significance of linguistic considerations in LLM pruning and advocates for language-specific optimization, establishing a framework for more efficient and tailored language models across diverse linguistic contexts. Additionally, all experiments were conducted using a single consumer-grade NVIDIA RTX 3090 GPU, and the code is available at https://github.com/mshamrai/language-specific-pruning.