Complex Word Identification (CWI) is an essential step in the lexical simplification task and has recently become a task on its own. Some variations of this binary classification task have emerged, such as lexical complexity prediction (LCP) and complexity evaluation of multi-word expressions (MWE). Large language models (LLMs) recently became popular in the Natural Language Processing community because of their versatility and capability to solve unseen tasks in zero/few-shot settings. Our work investigates LLM usage, specifically open-source models such as Llama 2, Llama 3, and Vicuna v1.5, and closed-source, such as ChatGPT-3.5-turbo and GPT-4o, in the CWI, LCP, and MWE settings. We evaluate zero-shot, few-shot, and fine-tuning settings and show that LLMs struggle in certain conditions or achieve comparable results against existing methods. In addition, we provide some views on meta-learning combined with prompt learning. In the end, we conclude that the current state of LLMs cannot or barely outperform existing methods, which are usually much smaller.

The remarkable achievements obtained by open-source large language models (LLMs) in recent years have predominantly been concentrated on tasks involving the English language. In this paper, we aim to advance the performance of Llama2 models on Romanian tasks. We tackle the problem of reduced computing resources by using QLoRA for training. We release RoQLlama-7b, a quantized LLM, which shows equal or improved results compared to its full-sized counterpart when tested on seven Romanian downstream tasks in the zero-shot setup. Also, it consistently achieves higher average scores across all few-shot prompts. Additionally, we introduce a novel Romanian dataset, namely RoMedQA, which contains single-choice medical questions in Romanian.

Multiword expressions are a key ingredient for developing large-scale and linguistically sound natural language processing technology. This paper describes our improvements in automatically identifying Romanian multiword expressions on the corpus released for the PARSEME v1.2 shared task. Our approach assumes a multilingual perspective based on the recently introduced lateral inhibition layer and adversarial training to boost the performance of the employed multilingual language models. With the help of these two methods, we improve the F1-score of XLM-RoBERTa by approximately 2.7% on unseen multiword expressions, the main task of the PARSEME 1.2 edition. In addition, our results can be considered SOTA performance, as they outperform the previous results on Romanian obtained by the participants in this competition.

Unsupervised Domain Adaptation (UDA) is a popular technique that aims to reduce the domain shift between two data distributions. It was successfully applied in computer vision and natural language processing. In the current work, we explore the effects of various unsupervised domain adaptation techniques between two text classification tasks: fake and hyperpartisan news detection. We investigate the knowledge transfer from fake to hyperpartisan news detection without involving target labels during training. Thus, we evaluate UDA, cluster alignment with a teacher, and cross-domain contrastive learning. Extensive experiments show that these techniques improve performance, while including data augmentation further enhances the results. In addition, we combine clustering and topic modeling algorithms with UDA, resulting in improved performances compared to the initial UDA setup.

Running large-scale pre-trained language models in computationally constrained environments remains a challenging problem yet to be addressed, while transfer learning from these models has become prevalent in Natural Language Processing tasks. Several solutions, including knowledge distillation, network quantization, or network pruning have been previously proposed; however, these approaches focus mostly on the English language, thus widening the gap when considering low-resource languages. In this work, we introduce three light and fast versions of distilled BERT models for the Romanian language: Distil-BERT-base-ro, Distil-RoBERT-base, and DistilMulti-BERT-base-ro. The first two models resulted from the individual distillation of knowledge from two base versions of Romanian BERTs available in literature, while the last one was obtained by distilling their ensemble. To our knowledge, this is the first attempt to create publicly available Romanian distilled BERT models, which were thoroughly evaluated on five tasks: part-of-speech tagging, named entity recognition, sentiment analysis, semantic textual similarity, and dialect identification. Our experimental results argue that the three distilled models offer performance comparable to their teachers, while being twice as fast on a GPU and ~35% smaller. In addition, we further test the similarity between the predictions of our students versus their teachers by measuring their label and probability loyalty, together with regression loyalty - a new metric introduced in this work.

In recent times, the detection of hate-speech, offensive, or abusive language in online media has become an important topic in NLP research due to the exponential growth of social media and the propagation of such messages, as well as their impact. Misogyny detection, even though it plays an important part in hate-speech detection, has not received the same attention. In this paper, we describe our classification systems submitted to the SemEval-2022 Task 5: MAMI - Multimedia Automatic Misogyny Identification. The shared task aimed to identify misogynous content in a multi-modal setting by analysing meme images together with their textual captions. To this end, we propose two models based on the pre-trained UNITER model, one enhanced with an image sentiment classifier, whereas the second leverages a Vocabulary Graph Convolutional Network (VGCN). Additionally, we explore an ensemble using the aforementioned models. Our best model reaches an F1-score of 71.4% in Sub-task A and 67.3% for Sub-task B positioning our team in the upper third of the leaderboard. We release the code and experiments for our models on GitHub.

Named Entity Recognition (NER) is a well-explored area from Information Retrieval and Natural Language Processing with an extensive research community. Despite that, few languages, such as English and German, are well-resourced, whereas many other languages, such as Romanian, have scarce resources, especially in domain-specific applications. In this work, we address the NER problem in the legal domain from both Romanian and German languages and evaluate the performance of our proposed method based on domain adaptation. We employ multi-task learning to jointly train a neural network on two legal and general domains and perform adaptation among them. The results show that domain adaptation increase performances by a small amount, under 1%, while considerable improvements are in the recall metric.

The real-world impact of polarization and toxicity in the online sphere marked the end of 2020 and the beginning of this year in a negative way. Semeval-2021, Task 5 - Toxic Spans Detection is based on a novel annotation of a subset of the Jigsaw Unintended Bias dataset and is the first language toxicity detection task dedicated to identifying the toxicity-level spans. For this task, participants had to automatically detect character spans in short comments that render the message as toxic. Our model considers applying Virtual Adversarial Training in a semi-supervised setting during the fine-tuning process of several Transformer-based models (i.e., BERT and RoBERTa), in combination with Conditional Random Fields. Our approach leads to performance improvements and more robust models, enabling us to achieve an F1-score of 65.73% in the official submission and an F1-score of 66.13% after further tuning during post-evaluation.

Extracting semantic information on measurements and counts is an important topic in terms of analyzing scientific discourses. The 8th task of SemEval-2021: Counts and Measurements (MeasEval) aimed to boost research in this direction by providing a new dataset on which participants train their models to extract meaningful information on measurements from scientific texts. The competition is composed of five subtasks that build on top of each other: (1) quantity span identification, (2) unit extraction from the identified quantities and their value modifier classification, (3) span identification for measured entities and measured properties, (4) qualifier span identification, and (5) relation extraction between the identified quantities, measured entities, measured properties, and qualifiers. We approached these challenges by first identifying the quantities, extracting their units of measurement, classifying them with corresponding modifiers, and afterwards using them to jointly solve the last three subtasks in a multi-turn question answering manner. Our best performing model obtained an overlapping F1-score of 36.91% on the test set.

Reading is a complex process which requires proper understanding of texts in order to create coherent mental representations. However, comprehension problems may arise due to hard-to-understand sections, which can prove troublesome for readers, while accounting for their specific language skills. As such, steps towards simplifying these sections can be performed, by accurately identifying and evaluating difficult structures. In this paper, we describe our approach for the SemEval-2021 Task 1: Lexical Complexity Prediction competition that consists of a mixture of advanced NLP techniques, namely Transformer-based language models, pre-trained word embeddings, Graph Convolutional Networks, Capsule Networks, as well as a series of hand-crafted textual complexity features. Our models are applicable on both subtasks and achieve good performance results, with a MAE below 0.07 and a Person correlation of .73 for single word identification, as well as a MAE below 0.08 and a Person correlation of .79 for multiple word targets. Our results are just 5.46% and 6.5% lower than the top scores obtained in the competition on the first and the second subtasks, respectively.

Detecting humor is a challenging task since words might share multiple valences and, depending on the context, the same words can be even used in offensive expressions. Neural network architectures based on Transformer obtain state-of-the-art results on several Natural Language Processing tasks, especially text classification. Adversarial learning, combined with other techniques such as multi-task learning, aids neural models learn the intrinsic properties of data. In this work, we describe our adversarial multi-task network, AMTL-Humor, used to detect and rate humor and offensive texts from Task 7 at SemEval-2021. Each branch from the model is focused on solving a related task, and consists of a BiLSTM layer followed by Capsule layers, on top of BERTweet used for generating contextualized embeddings. Our best model consists of an ensemble of all tested configurations, and achieves a 95.66% F1-score and 94.70% accuracy for Task 1a, while obtaining RMSE scores of 0.6200 and 0.5318 for Tasks 1b and 2, respectively.

This paper presents our contribution to the Social Media Mining for Health Applications Shared Task 2021. We addressed all the three subtasks of Task 1: Subtask A (classification of tweets containing adverse effects), Subtask B (extraction of text spans containing adverse effects) and Subtask C (adverse effects resolution). We explored various pre-trained transformer-based language models and we focused on a multi-task training architecture. For the first subtask, we also applied adversarial augmentation techniques and we formed model ensembles in order to improve the robustness of the prediction. Our system ranked first at Subtask B with 0.51 F1 score, 0.514 precision and 0.514 recall. For Subtask A we obtained 0.44 F1 score, 0.49 precision and 0.39 recall and for Subtask C we obtained 0.16 F1 score with 0.16 precision and 0.17 recall.

Dialect identification is a task with applicability in a vast array of domains, ranging from automatic speech recognition to opinion mining. This work presents our architectures used for the VarDial 2021 Romanian Dialect Identification subtask. We introduced a series of solutions based on Romanian or multilingual Transformers, as well as adversarial training techniques. At the same time, we experimented with a knowledge distillation tool in order to check whether a smaller model can maintain the performance of our best approach. Our best solution managed to obtain a weighted F1-score of 0.7324, allowing us to obtain the 2nd place on the leaderboard.

Dialect identification represents a key aspect for improving a series of tasks, for example, opinion mining, considering that the location of the speaker can greatly influence the attitude towards a subject. In this work, we describe the systems developed by our team for VarDial 2020: Romanian Dialect Identification, a task specifically created for challenging participants to solve the previously mentioned issue. More specifically, we introduce a series of neural systems based on Transformers, that combine a BERT model exclusively pre-trained on the Romanian language with techniques such as adversarial training or character-level embeddings. By using these approaches, we were able to obtain a 0.6475 macro F1 score on the test dataset, thus allowing us to be ranked 5th out of 8 participant teams.

This work presents our contribution in the context of the 6th task of SemEval-2020: Extracting Definitions from Free Text in Textbooks (DeftEval). This competition consists of three subtasks with different levels of granularity: (1) classification of sentences as definitional or non-definitional, (2) labeling of definitional sentences, and (3) relation classification. We use various pretrained language models (i.e., BERT, XLNet, RoBERTa, SciBERT, and ALBERT) to solve each of the three subtasks of the competition. Specifically, for each language model variant, we experiment by both freezing its weights and fine-tuning them. We also explore a multi-task architecture that was trained to jointly predict the outputs for the second and the third subtasks. Our best performing model evaluated on the DeftEval dataset obtains the 32nd place for the first subtask and the 37th place for the second subtask. The code is available for further research at: https://github.com/avramandrei/DeftEval

Users from the online environment can create different ways of expressing their thoughts, opinions, or conception of amusement. Internet memes were created specifically for these situations. Their main purpose is to transmit ideas by using combinations of images and texts such that they will create a certain state for the receptor, depending on the message the meme has to send. These posts can be related to various situations or events, thus adding a funny side to any circumstance our world is situated in. In this paper, we describe the system developed by our team for SemEval-2020 Task 8: Memotion Analysis. More specifically, we introduce a novel system to analyze these posts, a multimodal multi-task learning architecture that combines ALBERT for text encoding with VGG-16 for image representation. In this manner, we show that the information behind them can be properly revealed. Our approach achieves good performance on each of the three subtasks of the current competition, ranking 11th for Subtask A (0.3453 macro F1-score), 1st for Subtask B (0.5183 macro F1-score), and 3rd for Subtask C (0.3171 macro F1-score) while exceeding the official baseline results by high margins.

Sentiment analysis is a process widely used in opinion mining campaigns conducted today. This phenomenon presents applications in a variety of fields, especially in collecting information related to the attitude or satisfaction of users concerning a particular subject. However, the task of managing such a process becomes noticeably more difficult when it is applied in cultures that tend to combine two languages in order to express ideas and thoughts. By interleaving words from two languages, the user can express with ease, but at the cost of making the text far less intelligible for those who are not familiar with this technique, but also for standard opinion mining algorithms. In this paper, we describe the systems developed by our team for SemEval-2020 Task 9 that aims to cover two well-known code-mixed languages: Hindi-English and Spanish-English. We intend to solve this issue by introducing a solution that takes advantage of several neural network approaches, as well as pre-trained word embeddings. Our approach (multlingual BERT) achieves promising performance on the Hindi-English task, with an average F1-score of 0.6850, registered on the competition leaderboard, ranking our team 16 out of 62 participants. For the Spanish-English task, we obtained an average F1-score of 0.7064 ranking our team 17th out of 29 participants by using another multilingual Transformer-based model, XLM-RoBERTa.

Manipulative and misleading news have become a commodity for some online news outlets and these news have gained a significant impact on the global mindset of people. Propaganda is a frequently employed manipulation method having as goal to influence readers by spreading ideas meant to distort or manipulate their opinions. This paper describes our participation in the SemEval-2020, Task 11: Detection of PropagandaTechniques in News Articles competition. Our approach considers specializing a pre-trained BERT model on propagandistic and hyperpartisan news articles, enabling it to create more adequate representations for the two subtasks, namely propaganda Span Identification (SI) and propaganda Technique Classification (TC). Our proposed system achieved a F1-score of 46.060% in subtask SI, ranking 5th in the leaderboard from 36 teams and a micro-averaged F1 score of 54.302% for subtask TC, ranking 19th from 32 teams.

Offensive language detection is one of the most challenging problem in the natural language processing field, being imposed by the rising presence of this phenomenon in online social media. This paper describes our Transformer-based solutions for identifying offensive language on Twitter in five languages (i.e., English, Arabic, Danish, Greek, and Turkish), which was employed in Subtask A of the Offenseval 2020 shared task. Several neural architectures (i.e., BERT, mBERT, Roberta, XLM-Roberta, and ALBERT), pre-trained using both single-language and multilingual corpora, were fine-tuned and compared using multiple combinations of datasets. Finally, the highest-scoring models were used for our submissions in the competition, which ranked our team 21st of 85, 28th of 53, 19th of 39, 16th of 37, and 10th of 46 for English, Arabic, Danish, Greek, and Turkish, respectively.

Financial causality detection is centered on identifying connections between different assets from financial news in order to improve trading strategies. FinCausal 2020 - Causality Identification in Financial Documents – is a competition targeting to boost results in financial causality by obtaining an explanation of how different individual events or chain of events interact and generate subsequent events in a financial environment. The competition is divided into two tasks: (a) a binary classification task for determining whether sentences are causal or not, and (b) a sequence labeling task aimed at identifying elements related to cause and effect. Various Transformer-based language models were fine-tuned for the first task and we obtained the second place in the competition with an F1-score of 97.55% using an ensemble of five such language models. Subsequently, a BERT model was fine-tuned for the second task and a Conditional Random Field model was used on top of the generated language features; the system managed to identify the cause and effect relationships with an F1-score of 73.10%. We open-sourced the code and made it available at: https://github.com/avramandrei/FinCausal2020.

This paper describes our solutions submitted to the Social Media Mining for Health Applications (#SMM4H) Shared Task 2020. We participated in the following tasks: Task 1 aimed at classifying if a tweet reports medications or not, Task 2 (only for the English dataset) aimed at discriminating if a tweet mentions adverse effects or not, and Task 5 aimed at recognizing if a tweet mentions birth defects or not. Our work focused on studying different neural network architectures based on various flavors of bidirectional Transformers (i.e., BERT), in the context of the previously mentioned classification tasks. For Task 1, we achieved an F1-score (70.5%) above the mean performance of the best scores made by all teams, whereas for Task 2, we obtained an F1-score of 37%. Also, we achieved a micro-averaged F1-score of 62% for Task 5.

In recent years, the need for communication increased in online social media. Propaganda is a mechanism which was used throughout history to influence public opinion and it is gaining a new dimension with the rising interest of online social media. This paper presents our submission to NLP4IF-2019 Shared Task SLC: Sentence-level Propaganda Detection in news articles. The challenge of this task is to build a robust binary classifier able to provide corresponding propaganda labels, propaganda or non-propaganda. Our model relies on a unified neural network, which consists of several deep leaning modules, namely BERT, BiLSTM and Capsule, to solve the sentencelevel propaganda classification problem. In addition, we take a pre-training approach on a somewhat similar task (i.e., emotion classification) improving results against the cold-start model. Among the 26 participant teams in the NLP4IF-2019 Task SLC, our solution ranked 12th with an F1-score 0.5868 on the official test data. Our proposed solution indicates promising results since our system significantly exceeds the baseline approach of the organizers by 0.1521 and is slightly lower than the winning system by 0.0454.

This paper describes our models for the Moldavian vs. Romanian Cross-Topic Identification (MRC) evaluation campaign, part of the VarDial 2019 workshop. We focus on the three subtasks for MRC: binary classification between the Moldavian (MD) and the Romanian (RO) dialects and two cross-dialect multi-class classification between six news topics, MD to RO and RO to MD. We propose several deep learning models based on long short-term memory cells, Bidirectional Gated Recurrent Unit (BiGRU) and Hierarchical Attention Networks (HAN). We also employ three word embedding models to represent the text as a low dimensional vector. Our official submission includes two runs of the BiGRU and HAN models for each of the three subtasks. The best submitted model obtained the following macro-averaged F1 scores: 0.708 for subtask 1, 0.481 for subtask 2 and 0.480 for the last one. Due to a read error caused by the quoting behaviour over the test file, our final submissions contained a smaller number of items than expected. More than 50% of the submission files were corrupted. Thus, we also present the results obtained with the corrected labels for which the HAN model achieves the following results: 0.930 for subtask 1, 0.590 for subtask 2 and 0.687 for the third one.

Understanding questions and answers in QA system is a major challenge in the domain of natural language processing. In this paper, we present a question answering system that influences the human opinions in a conversation. The opinion words are quantified by using a lexicon-based method. We apply Latent Semantic Analysis and the cosine similarity measure between candidate answers and each question to infer the answer of the chatbot.