Search

Background: The pandemic of the coronavirus disease 2019 (COVID-19) has led to an increased burden on mental health. Aims: To investigate the development of major depressive disorder (MDD), generalized anxiety disorder (GAD), and suicidal ideation in the Netherlands during the first fifteen months of the pandemic and three nation-wide lockdowns. Method: Participants of the Lifelines Cohort Study –a Dutch population-based sample-reported current symptoms of MDD and GAD, including suicidal ideation, according to DSM-IV criteria. Between March 2020 and June 2021, 36,106 participants (aged 18–96) filled out a total of 629,811 questionnaires across 23 time points. Trajectories over time were estimated using generalized additive models and analyzed in relation to age, sex, and lifetime history of MDD/GAD. Results: We found non-linear trajectories for MDD and GAD with a higher number of symptoms and prevalence rates during periods of lockdown. The point prevalence of MDD and GAD peaked during the third hard lockdown at 2.88 % (95 % CI: 2.71 %–3.06 %) and 2.92 % (95 % CI: 2.76 %–3.08 %), respectively, in March 2021. Women, younger adults, and participants with a history of MDD/GAD reported significantly more symptoms. For suicidal ideation, we found a significant linear increase over time in younger participants. For example, 20-year-old participants reported 4.14× more suicidal ideation at the end of June 2021 compared to the start of the pandemic (4.64 % (CI: 3.09 %–6.96 %) versus 1.12 % (CI: 0.76 %–1.66 %)). Limitations: Our findings should be interpreted in relation to the societal context of the Netherlands and the public health response of the Dutch government during the pandemic, which may be different in other regions in the world. Conclusions: Our study showed greater prevalence of MDD and GAD during COVID-19 lockdowns and a continuing increase in suicidal thoughts among young adults suggesting that the pandemic and government enacted restrictions impacted mental health in the population. Our findings provide actionable insights on mental health in the population during the pandemic, which can guide policy makers and clinical care during future lockdowns and epi/pandemics.

The present study examined the kinematics of maximal effort sprint running, mapping the relations among a person's maximal running speed, maximum running acceleration and the distance coverable in a certain amount of time by this person. Thirty-three participants were recruited to perform a simple sprint task. Both forward and backward running were considered. Participants' position, velocity and acceleration data were obtained using a Local Positioning Measurement system. Participants' speed-acceleration profiles turned out to be markedly non-linear. To account for these non-linear patterns, we propose a new macroscopic model on the kinematics of sprint running. Second, we examined whether target distance was of influence on the evolution of participants' running speeds over time. Overall, no such effect on running velocity was present, except for a 'finish-line effect'. Finally, we studied how variation in individuals' maximum running velocities and accelerations related to differences in their action boundaries. The findings are discussed in the context of affordance-based control in running to catch fly balls.

Although communicative language teaching (CLT) was thought to have revolutionized classroom practice, there are “weak” and “strong” versions (Howatt, 1984). Most foreign language classrooms in the world still favor weak versions with structure-based (SB) views on language (Lightbown & Spada, 2013), and practice in the Netherlands is not much different (West & Verspoor, 2016). However, a small group of teachers in the Netherlands started teaching French as a second language with a strong CLT program in line with Dynamic Usage-Based (DUB) principles. Rather than focusing on rule learning and explicit grammar teaching to avoid errors, the DUB program takes the dynamics of second-language development into consideration and focuses on the three key elements of usage-based theory: frequency, salience and contingency. These translate into a great deal of exposure, repetition, learning the meaning of every single word through gestures, and presenting whole chunks of language, all without explicit grammar teaching. This study aims to compare the effects of the SB and DUB instructional programs after three years. We traced the second-language development of 229 junior high school students (aged 12 to 15) learning French in the Netherlands over three years. The participants took three oral tests over the course of three years (568 interviews) and wrote seven narratives on the same topic (1511 narratives). As expected, the DUB approach, which is in line with a strong CLT version, was more effective in achieving proficiency in both speaking and writing and equally effective in achieving accuracy.

In this paper, we present an in-depth investigation of the linguistic knowledge encoded by the transformer models currently available for the Italian language. In particular, we investigate how the complexity of two different architectures of probing models affects the performance of the Transformers in encoding a wide spectrum of linguistic features. Moreover, we explore how this implicit knowledge varies according to different textual genres and language varieties.

This article presents the results of a study involving the translation of a short story by Kurt Vonnegut from English to Catalan and Dutch using three modalities: machine-translation (MT), post-editing (PE) and translation without aid (HT). Our aim is to explore creativity, understood to involve novelty and acceptability, from a quantitative perspective. The results show that HT has the highest creativity score, followed by PE, and lastly, MT, and this is unanimous from all reviewers. A neural MT system trained on literary data does not currently have the necessary capabilities for a creative translation; it renders literal solutions to translation problems. More importantly, using MT to post-edit raw output constrains the creativity of translators, resulting in a poorer translation often not fit for publication, according to experts., 28 pages, 2 figures, 10 tables

Studies on the applicability of heterogeneous semantically interoperable corpora are rare. We investigate to what extent reusability (both of systems and of annotations) is entailed by corpora whose interoperability is based on compliance to standards. In particular, we look at event detection in English texts, supported by the ISO-TimeML annotation scheme. We run two sets of experiments using a common neural network architecture and extensively evaluate our results on both in-distribution and out-of-distribution settings. In all experimental settings, systems obtain state-of-the-art results on the in-distribution data and underperform out-of-distribution ones, setting limits to the benefits of semantically interoperable corpora. By means of a detailed error analysis, we show that while being compliant to a standard guarantees semantic interoperability, this becomes only a necessary condition for reusability, with factors such as differences in the quality of the annotations having a much stronger impact.

This study aims to show how a museum exhibition designed to encourage visitors to take pictures can affect visitors’ behavior. We analyzed visitors’ engagement with the Yumi's Cell Special Exhibition (hereafter Yumi) held in South Korea, employing computer vision for the analysis of visitors’ Instagram pictures. Our research questions are: What types of pictures do visitors post on Instagram during or after their visit? And can Yumi’s instagrammable features make visitors interact more with the exhibition? We also formulated two corresponding hypotheses: Visitors are primarily interested in taking selfies in the instagrammable environment; and visitors struck more active poses when taking pictures in an instagrammable exhibition than in a traditional art exhibition. Through the image analysis of Instagram posts of the exhibition, we found many pictures of people, but the proportion of selfies was relatively limited. This suggests that visitors were more interested in interacting with the exhibition rather than taking selfies. This has also been confirmed by the pose analysis, which showed that the participatory feature of the exhibition encouraged visitors to take photos in active poses, interacting, mimicking, and performing. The framework presented and the findings offer insights about how to design exhibitions to increase visitors’ participation.

Previous experimental findings support the hypothesis that laughter and positive emotions are contagious in face-to-face and mediated communication. To test this hypothesis, we describe four experiments in which participants communicate via a chat tool that artificially adds or removes laughter (e.g. haha or lol), without participants being aware of the manipulation. We found no evidence to support the contagion hypothesis. However, artificially exposing participants to more lols decreased participants’ use of hahas but led to more involvement and improved task-performance. Similarly, artificially exposing participants to more hahas decreased use of haha but increased lexical alignment. We conclude that, even though the interventions have effects on coordination, they are incompatible with contagion as a primary explanatory mechanism. Instead, these results point to an interpretation that involves a more sophisticated view of dialogue mechanisms along the lines of Conversational Analysis and similar frameworks and we suggest directions for future research.

The first workshop on Argumentation Knowledge Graphs (ArgKG) was held virtually at the Automated Knowledge Base Construction (AKBC 2021) conference on October 7, 2021. ArgKG @ AKBC 2021 brought together the Computational Argumentation and Knowledge Graphs communities, aiming to promote cross-pollination of ideas and encourage discussions and collaborations between the two communities. This paper describes the workshop and compiles several of its findings and insights. Date : 7 October, 2021. Website : https://argkg21.argmining.org.

Introduction: Surgical treatment of oral cancer leads to lasting changes of the vocal tract and individuals treated for oral cancer (ITFOC) often experience speech problems. The purpose of this study was to analyse the acoustic properties of the spontaneous speech of individuals who were surgically treated for oral cancer. It was investigated (1) how key spectral measures of articulation change post-treatment; (2) whether changes are more related to target manner or place of articulation; and (3) how spectral measures develop at various time points following treatment.Method: A corpus consisting of 32.850 tokens was constructed by manually segmenting the speech of five (four female - one male) American English speaking ITFOC. General acoustic characteristics (duration and spectral tilt), plosives (burst frequency), fricatives (centre of gravity and spectral skewness), and vowels (F1 and F2) were analysed using linear mixed effects regression and compared to control speech. Moreover, a within speaker analysis was performed for speakers with multiple recordings.Results: Manner of articulation is more predictive of post-treatment changes than place of articulation. Compared to controls, ITFOC produced the fricatives /f, v, θ, ð, s, z, ʃ, ʒ/ with a lower centre of gravity while no differences were found in plosives and vowels. Longitudinal analyses show high within-speaker variation, but general improvements one-year post-treatment.Conclusions: Surgical oral cancer treatment changes the spectral properties of speech. Fricatives with varying manner of articulations were distorted, suggesting that manner of articulation is more predictive than place of articulation in identifying general problem areas for ITFOC.

We present OpenBoek: a corpus of 103k tokens of classic Dutch novels with annotated coreference and entities. The corpus has several properties that are challenging for current coreference models: long documents (fragments of 10k+ words each), domain-specific literary phenomena, and 19th century Dutch spelling. Spelling normalization is added to the corpus as an additional annotation layer, using a data-driven rule-based spelling normalization tool. Normalizations are added using meta-annotation, such that evaluation can be performed with annotations on the original texts without losing token alignment. This tool enables the application of parsing and coreference systems originally developed for modern Dutch. We evaluate parsing and coreference systems on the OpenBoek dataset and find that spelling normalization gives a substantial increase in performance. The OpenBoek corpus is available under an open licens at https://andreasvc.github.io/openboek/

We introduce DivEMT, the first publicly available post-editing study of Neural Machine Translation (NMT) over a typologically diverse set of target languages. Using a strictly controlled setup, 18 professional translators were instructed to translate or post-edit the same set of English documents into Arabic, Dutch, Italian, Turkish, Ukrainian, and Vietnamese. During the process, their edits, keystrokes, editing times and pauses were recorded, enabling an in-depth, cross-lingual evaluation of NMT quality and post-editing effectiveness. Using this new dataset, we assess the impact of two state-of-the-art NMT systems, Google Translate and the multilingual mBART-50 model, on translation productivity. We find that post-editing is consistently faster than translation from scratch. However, the magnitude of productivity gains varies widely across systems and languages, highlighting major disparities in post-editing effectiveness for languages at different degrees of typological relatedness to English, even when controlling for system architecture and training data size. We publicly release the complete dataset including all collected behavioral data, to foster new research on the translation capabilities of NMT systems for typologically diverse languages.

Different linguistic expressions can conceptualize the same event from different viewpoints by emphasizing certain participants over others. Here, we investigate a case where this has social consequences: how do linguistic expressions of gender-based violence (GBV) influence who we perceive as responsible? We build on previous psycholinguistic research in this area and conduct a large-scale perception survey of GBV descriptions automatically extracted from a corpus of Italian newspapers. We then train regression models that predict the salience of GBV participants with respect to different dimensions of perceived responsibility. Our best model (fine-tuned BERT) shows solid overall performance, with large differences between dimensions and participants: salient _focus_ is more predictable than salient _blame_, and perpetrators' salience is more predictable than victims' salience. Experiments with ridge regression models using different representations show that features based on linguistic theory similarly to word-based features. Overall, we show that different linguistic choices do trigger different perceptions of responsibility, and that such perceptions can be modelled automatically. This work can be a core instrument to raise awareness of the consequences of different perspectivizations in the general public and in news producers alike., Accepted for publication at AACL-IJCNLP 2022

Most research on natural language processing treats bias as an absolute concept: Based on a (probably complex) algorithmic analysis, a sentence, an article, or a text is classified as biased or not. Given the fact that for humans the question of whether a text is biased can be difficult to answer or is answered contradictory, we ask whether an "absolute bias classification" is a promising goal at all. We see the problem not in the complexity of interpreting language phenomena but in the diversity of sociocultural backgrounds of the readers, which cannot be handled uniformly: To decide whether a text has crossed the proverbial line between non-biased and biased is subjective. By asking "Is text X more [less, equally] biased than text Y?" we propose to analyze a simpler problem, which, by its construction, is rather independent of standpoints, views, or sociocultural aspects. In such a model, bias becomes a preference relation that induces a partial ordering from least biased to most biased texts without requiring a decision on where to draw the line. A prerequisite for this kind of bias model is the ability of humans to perceive relative bias differences in the first place. In our research, we selected a specific type of bias in argumentation, the stance bias, and designed a crowdsourcing study showing that differences in stance bias are perceptible when (light) support is provided through training or visual aid.

SPOD is part of the PaQu website created as a CLARIN project. It allows one to generate a syntactic profile of a corpus based on the output of the automatic parser Alpino. It runs a long sequence of queries and provides quantitative information about constituents, sentence types, coordination, length of constituents, and so on. In this chapter, we employ SPOD and the rest of PaQu to analyse a part of the Schrijfmeterscorpus of secondary school essays. We use a small subsection of the SPOD output for this purpose, in particular those syntactic properties that correlate most reliably with academically oriented texts. We show that SPOD is able to distinguish, on the basis of these variables, among grades and school types.

Most of the work on Character Networks to date is limited to monolingual texts. Conversely, in this paper we apply and analyze Character Networks on both source texts (English novels) and their Finnish translations (both human- and machine-translated). We assume that this analysis could provide some insights on changes in translations that could modify the character networks, as well as the narrative. The results show that the character networks of translations differ from originals in case of long novels, and the differences may also vary depending on the novel and translator's strategy.

Even though many recent semantic parsers are based on deep learning methods, we should not forget that rule-based alternatives might offer advantages over neural approaches with respect to transparency, portability, and explainability. Taking advantage of existing off-the-shelf Universal Dependency parsers, we present a method that maps a syntactic dependency tree to a formal meaning representation based on Discourse Representation Theory. Rather than using lambda calculus to manage variable bindings, our approach is novel in that it consists of using a series of graph transformations. The resulting UD semantic parser shows good performance for English, German, Italian and Dutch, with F-scores over 75%, outperforming a neural semantic parser for the lower-resourced languages. Unlike neural semantic parsers, our UD semantic parser does not hallucinate output, is relatively easy to port to other languages, and is completely transparent.

Existing studies on the analysis of persuasion in online discussions focus on investigating the effectiveness of comments in discussions and ignore the analysis of the effectiveness of debaters over multiple discussions. In this paper, we propose to quantify debaters effectiveness in the online discussion platform: “ChangeMyView” in order to explore diverse insights into their persuasion strategies. In particular, targeting debaters with different levels of effectiveness (e.g., good vs. bad), various behavioral characteristics (e..g, engagement) and text stylistic features (e.g., used frames) of debaters are carefully examined, leading to several outcomes that can be the backbone of writing assistants and persuasive text generation.

We pursue the hypothesis that musical differences between cultures are based on linguistic, especially phonological, properties of the culture's spoken language. To study this hypothesis, we present a general constraint-based framework for describing the structural similarities between music and language. Music and language are structured by the fact that some sounds are more important than others, based on cognitive strategies which we present here as universal well-formedness conditions. However, which sounds are considered to be most salient differs across cultures, as evidenced by the world's many linguistic and musical typologies. The first goal of our research approach is to identify these universal well-formedness conditions (e.g. prominence of strong elements based on the syllable/chord structure and domain marking based on intonation/ melody patterns, pauses) for speech and music. The second goal is to assess how cultures differ from each other in terms of the relative salience assigned to these conditions (i.e. how these conditions are "ranked"). The current paper is meant to be an introduction to a new approach with focus on the identification of general well-formedness conditions. We introduce similar conditions for the description of language and music in order to make comparison of the two disciplines more fruitful. The goal of our research approach is to create a theoretical and methodological map to aid more detailed culture-specific comparisons. The ultimate aim is to provide a comprehensive typological overview for which we will start with a selection of culture families following the World Atlas of Language Structures online (Dryer & Haspelmath, 2013) for language and the Global Jukebox (Wood & Arèvalo, 2018) for music.

Machine Translation has come a long way since the first “mechanical translation” experiments of the post-war period, an era that also marked the birth of Translation Studies. In this introduction, we trace the histories of both areas of study, which are interconnected in many ways, even though they have often been perceived as opposite camps. In the second decade of the twenty-first century, Machine Translation and Translation Studies have started to converge. An area in which this convergence can be seen is in the application of Machine Translation (and other translation technologies) to the translation of creative texts, together with the study of the implications thereof, e.g., societal, ethical, and legal. This book is the result of cross-fertilisation between scholars in Machine Translation and Translation Studies, and Literary Studies. It is the first of its kind in creating a snapshot of the state of the art in the use of technology in translating creative texts.

This chapter discusses the application of neural machine translation (NMT) to works of fiction that have been published as e-books in a bilingual setting. These bilingual e-books are translated using NMT for readers with an intermediate or high level of a foreign language who want to read novels originally written in English. NMT is used to provide a translation in their mother tongue that the readers can refer to on demand, e.g., to help them understand difficult passages in the original. The main hypothesis of this research is that the quality provided by tailored NMT systems is sufficient to meet this goal and that using this methodology will enhance the readers’ experience. We expect the users of these bilingual e-books to read most of the book in the original language and to use the translation only sporadically when encountering difficult parts. This could help the reading experience become more intense, since the reader does not need to consult external sources such as dictionaries or use external public NMT engines.In this chapter, we first discuss our motivation and present the goals of the project. Then the previous work in this field is presented. After that, we explain the process of customizing an NMT engine and creating the bilingual e-books. We follow with the design of the survey and the results of a reading experiment using an English short story set up for Catalan native speakers with a between-subjects design. Finally, we reflect on the results and propose avenues for future work. We are mainly interested in describing a useful methodology for creating bilingual books with NMT and assessing their usability by involving L2 readers.

Deep acoustic models represent linguistic information based on massive amounts of data. Unfortunately, for regional languages and dialects such resources are mostly not available. However, deep acoustic models might have learned linguistic information that transfers to low-resource languages. In this study, we evaluate whether this is the case through the task of distinguishing low-resource (Dutch) regional varieties. By extracting embeddings from the hidden layers of various wav2vec 2.0 models (including new models which are pre-trained and/or fine-tuned on Dutch) and using dynamic time warping, we compute pairwise pronunciation differences averaged over 10 words for over 100 individual dialects from four (regional) languages. We then cluster the resulting difference matrix in four groups and compare these to a gold standard, and a partitioning on the basis of comparing phonetic transcriptions. Our results show that acoustic models outperform the (traditional) transcription-based approach without requiring phonetic transcriptions, with the best performance achieved by the multilingual XLSR-53 model fine-tuned on Dutch. On the basis of only six seconds of speech, the resulting clustering closely matches the gold standard., Accepted at NAACL 2022

This paper presents a comprehensive corpus for the study of socially unacceptable language in Dutch. The corpus extends and revise an existing resource with more data and introduces a new annotation dimension for offensive language, making it a unique resource in the Dutch language panorama. Each language phenomenon (abusive and offensive language) in the corpus has been annotated with a multilayer annotation scheme modelling the explicitness and the target(s) of the abuse/offence in the message. We have conducted a new set of experiments with different classification algorithms on all annotation dimensions. Monolingual Pre-Trained Language Models prove as the best systems, obtaining a macro-average F1 of 0.828 for binary classification of offensive language, and 0.579 for the targets of offensive messages. Furthermore, the best system obtains a macro-average F1 of 0.637 for distinguishing between abusive and offensive messages.

Character identification is a key element for many narrative-related tasks. To implement it, the baseform of the name of the character (or lemma) needs to be identified, so different appearances of the same character in the narrative could be aligned. In this paper we tackle this problem in translated texts (English--Finnish translation direction), where the challenge regarding lemmatizing foreign names in an agglutinative language appears. To solve this problem, we present and compare several methods. The results show that the method based on a search for the shortest version of the name proves to be the easiest, best performing (83.4% F1), and most resource-independent.

We present a lemmatizer/PoS tagger/dependency parser for West Frisian using a corpus of 44,714 words in 3,126 sentences that were annotated according to the guidelines of Universal Dependencies version 2. PoS tags were assigned to words by using a Dutch PoS tagger that was applied to a Dutch word-by-word translation, or to sentences of a Dutch parallel text. Best results were obtained when using word-by-word translations that were created by using the previous version of the Frisian translation program Oersetter. Morphologic and syntactic annotations were generated on the basis of a Dutch word-by-word translation as well. The performance of the lemmatizer/tagger/annotator when it was trained using default parameters was compared to the performance that was obtained when using the parameter values that were used for training the LassySmall UD 2.5 corpus. We study the effects of different hyperparameter settings on the accuracy of the annotation pipeline. The Frisian lemmatizer/PoS tagger/dependency parser is released as a web app and as a web service.

To facilitate corpus searches by classicists as well as to reduce data sparsity when training models, we focus on the automatic lemmatization of ancient Greek inscriptions, which have not received as much attention in this sense as literary text data has. We show that existing lemmatizers for ancient Greek, trained on literary data, are not performant on epigraphic data, due to major language differences between the two types of texts. We thus train the first inscription-specific lemmatizer achieving above 80% accuracy, and make both the models and the lemmatized data available to the community. We also provide a detailed error analysis highlighting peculiarities of inscriptions which again highlights the importance of a lemmatizer dedicated to inscriptions.

The scarcity of parallel data is a major limitation for Neural Machine Translation (NMT) systems, in particular for translation into morphologically rich languages (MRLs). An important way to overcome the lack of parallel data is to leverage target monolingual data, which is typically more abundant and easier to collect. We evaluate a number of techniques to achieve this, ranging from back-translation to random token masking, on the challenging task of translating English into four typologically diverse MRLs, under low-resource settings. Additionally, we introduce Inflection Pre-Training (or PT-Inflect), a novel pre-training objective whereby the NMT system is pre-trained on the task of re-inflecting lemmatized target sentences before being trained on standard source-to-target language translation. We conduct our evaluation on four typologically diverse target MRLs, and find that PT-Inflect surpasses NMT systems trained only on parallel data. While PT-Inflect is outperformed by back-translation overall, combining the two techniques leads to gains in some of the evaluated language pairs.

An important goal of the MaCoCu project is to improve EU-specific NLP systems that concern their Digital Service Infrastructures (DSIs). In this paper we aim at boosting the creation of such domain-specific NLP systems. To do so, we explore the feasibility of building an automatic classifier that allows to identify which segments in a generic (potentially parallel) corpus are relevant for a particular DSI. We create an evaluation data set by crawling DSI-specific web domains and then compare different strategies to build our DSI classifier for text in three languages: English, Spanish and Dutch. We use pre-trained (multilingual) language models to perform the classification, with zero-shot classification for Spanish and Dutch. The results are promising, as we are able to classify DSIs with between 70 and 80% accuracy, even without in-language training data. A manual annotation of the data revealed that we can also find DSI-specific data on crawled texts from general web domains with reasonable accuracy. We publicly release all data, predictions and code, as to allow future investigations in whether exploiting this DSI-specific data actually leads to improved performance on particular applications, such as machine translation.

Neural machine translation (NMT) systems are nowadays essential components of professional translation workflows. Consequently, human translators are increasingly working as post-editors for machine-translated content. The NWO-funded InDeep project aims to empower users of Deep Learning models of text, speech, and music by improving their ability to interact with such models and interpret their behaviors. In the specific context of translation, we aim at developing new tools and methodologies to improve prediction attribution, error analysis, and controllable generation for NMT systems. These advances will be evaluated through field studies involving professional translators to assess gains in efficiency and overall enjoyability of the post-editing process.

We present here the EU-funded project CREAMT that seeks to understand what is meant by creativity in different translation modalities, e.g. machine translation, post-editing or professional translation. Focusing on the textual elements that determine creativity in translated literary texts and the reader experience, CREAMT uses a novel, interdisciplinary approach to assess how effective MT is in literary translation considering creativity in translation and the ultimate user: the reader.

We address the task of automatically distinguishing between human-translated (HT) and machine translated (MT) texts. Following recent work, we fine-tune pre-trained language models (LMs) to perform this task. Our work differs in that we use state-of-the-art pre-trained LMs, as well as the test sets of the WMT news shared tasks as training data, to ensure the sentences were not seen during training of the MT system itself. Moreover, we analyse performance for a number of different experimental setups, such as adding translationese data, going beyond the sentence-level and normalizing punctuation. We show that (i) choosing a state-of-the-art LM can make quite a difference: our best baseline system (DeBERTa) outperforms both BERT and RoBERTa by over 3% accuracy, (ii) adding translationese data is only beneficial if there is not much data available, (iii) considerable improvements can be obtained by classifying at the document-level and (iv) normalizing punctuation and thus avoiding (some) shortcuts has no impact on model performance.

Dutchauthors have been criticizing the homogeneity and the dominance of white men authors in Dutch literary prize nominations and the Dutch literary scene.This homogeneity is clearly seen in the Dutch literary prizes, as 80% of the winners of the two most prominent Dutch literary prizes are men. This researchinvestigateswhether it is possible to identify author gender inequality in Dutch literary prizes using quantifiable literary qualities.&nbsp

IntroductionIt has been proposed that bilinguals’ language use patterns are differentially associated with executive control. To further examine this, the present study relates the social diversity of bilingual language use to performance on a color-shape switching task (CSST) in a group of bilingual university students with diverse linguistic backgrounds. Crucially, this study used language entropy as a measure of bilinguals’ language use patterns. This continuous measure reflects a spectrum of language use in a variety of social contexts, ranging from compartmentalized use to fully integrated use.MethodsLanguage entropy for university and non-university contexts was calculated from questionnaire data on language use. Reaction times (RTs) were measured to calculate global RT and switching and mixing costs on the CSST, representing conflict monitoring, mental set shifting, and goal maintenance, respectively. In addition, this study innovatively recorded a potentially more sensitive measure of set shifting abilities, namely, pupil size during task performance.ResultsHigher university entropy was related to slower global RT. Neither university entropy nor non-university entropy were associated with switching costs as manifested in RTs. However, bilinguals with more compartmentalized language use in non-university contexts showed a larger difference in pupil dilation for switch trials in comparison with non-switch trials. Mixing costs in RTs were reduced for bilinguals with higher diversity of language use in non-university contexts. No such effects were found for university entropy.DiscussionThese results point to the social diversity of bilinguals’ language use as being associated with executive control, but the direction of the effects may depend on social context (university vs. non-university). Importantly, the results also suggest that some of these effects may only be detected by using more sensitive measures, such as pupil dilation. The paper discusses theoretical and practical implications regarding the language entropy measure and the cognitive effects of bilingual experiences more generally, as well as how methodological choices can advance our understanding of these effects.

Variation in speech is often quantified by comparing phonetic transcriptions of the same utterance. However, manually transcribing speech is time-consuming and error prone. As an alternative, therefore, we investigate the extraction of acoustic embeddings from several self-supervised neural models. We use these representations to compute word-based pronunciation differences between non-native and native speakers of English, and between Norwegian dialect speakers. For comparison with several earlier studies, we evaluate how well these differences match human perception by comparing them with available human judgements of similarity. We show that speech representations extracted from a specific type of neural model (i.e. Transformers) lead to a better match with human perception than two earlier approaches on the basis of phonetic transcriptions and MFCC-based acoustic features. We furthermore find that features from the neural models can generally best be extracted from one of the middle hidden layers than from the final layer. We also demonstrate that neural speech representations not only capture segmental differences, but also intonational and durational differences that cannot adequately be represented by a set of discrete symbols used in phonetic transcriptions., Submitted to Journal of Phonetics

Identifying factors that make certain languages harder to model than others is essential to reach language equality in future Natural Language Processing technologies. Free-order case-marking languages, such as Russian, Latin or Tamil, have proved more challenging than fixed-order languages for the tasks of syntactic parsing and subject-verb agreement prediction. In this work, we investigate whether this class of languages is also more difficult to translate by state-of-the-art Neural Machine Translation models (NMT). Using a variety of synthetic languages and a newly introduced translation challenge set, we find that word order flexibility in the source language only leads to a very small loss of NMT quality, even though the core verb arguments become impossible to disambiguate in sentences without semantic cues. The latter issue is indeed solved by the addition of case marking. However, in medium- and low-resource settings, the overall NMT quality of fixed-order languages remains unmatched., Comment: Accepted to TACL, pre-MIT Press publication version

The digital preservation of funerary material is of interest to many different scientific disciplines. Textual information found on tombstones often goes far beyond the expected (name of the deceased, dates of birth and death), and may include information about commemorators, family roles, occupations, references to biblical or other texts, places of birth and death, cause of death, epitaphs and poems. Gravestones are multi-modal media, and besides text are often decorated with artistic symbols. To capture this information in a systematic way and make it available on a large scale for research purposes, a meaning representation based on linking entities by relations has been designed that will extend search capabilities beyond simple string matches. Concepts are represented as WordNet synsets, and a vocabulary of 32 relations make connections between concepts. This formalisation has been developed and evaluated based on a dataset of more than 1,000 Dutch tombstones.

The fifth edition of the CheckThat! Lab is held as part of the 2022 Conference and Labs of the Evaluation Forum (CLEF). The lab evaluates technology supporting various factuality tasks in seven languages: Arabic, Bulgarian, Dutch, English, German, Spanish, and Turkish. Task1 focuses on disinformation related to the ongoing COVID-19 infodemic and politics, and asks to predict whether a tweet is worth fact-checking, contains a verifiable factual claim, is harmful to the society, or is of interest to policy makers and why. Task2 asks to retrieve claims that have been previously fact-checked and that could be useful to verify the claim in a tweet. Task3 is to predict the veracity of a news article. Tasks1 and 3 are classification problems, while Task2 is a ranking one.

We introduce a framework for the computational analysis of how responsibility is framed in the reporting of two types of socially relevant events: gender-based violence (specifically, femicides in the Italian press), and traffic danger (specifically, traffic crashes in Dutch and Flemish news reports). We advocate for the parallel analysis of these two phenomena under the same theoretical framework, which draws on Frame Semantics, Critical Discourse Analysis and Natural Language Processing. Reusing two existing event-text datasets we show how computational experiments and the resulting analyses can be run. This work supports the testing and development of tools for NLP practitioners, as well as large-scale linguistic analyses for activists and journalists, in the context of socially impacting events.

Cross-lingual transfer learning with large multilingual pre-trained models can be an effective approach for low-resource languages with no labeled training data. Existing evaluations of zero-shot cross-lingual generalisability of large pre-trained models use datasets with English training data, and test data in a selection of target languages. We explore a more extensive transfer learning setup with 65 different source languages and 105 target languages for part-of-speech tagging. Through our analysis, we show that pre-training of both source and target language, as well as matching language families, writing systems, word order systems, and lexical-phonetic distance significantly impact cross-lingual performance. The findings described in this paper can be used as indicators of which factors are important for effective zero-shot cross-lingual transfer to zero- and low-resource languages.

Connotation is a dimension of lexical meaning at the semantic-pragmatic interface. Connotations can be used to express point of views, perspectives, and implied emotional associations. Variations in connotations of the same lexical item can occur at different level of analysis: from individuals, to community of speech, specific domains, and even time. In this paper, we present a simple yet effective method to assign connotative values to selected target items and to quantify connotation shifts. We test our method via a set of experiments using different social media data (Reddit and Twitter) and languages (English and Italian). While we kept the connotative axis (i.e., the polarity associated to a lexical item) fixed, we investigated connotation shifts along two dimensions: the first target shifts across communities of speech and domain while the second targets shifts in time. Our results indicate the validity of the proposed method and its potential application for the identification of connotation shifts and application to automatically induce specific connotation lexica.

Massively multilingual models are promising for transfer learning across tasks and languages. However, existing methods are unable to fully leverage training data when it is available in different task-language combinations. To exploit such heterogeneous supervision, we propose Hyper-X, a single hypernetwork that unifies multi-task and multilingual learning with efficient adaptation. It generates weights for adapter modules conditioned on both tasks and language embeddings. By learning to combine task and language-specific knowledge, our model enables zero-shot transfer for unseen languages and task-language combinations. Our experiments on a diverse set of languages demonstrate that Hyper-X achieves the best or competitive gain when a mixture of multiple resources is available, while on par with strong baseline in the standard scenario. Hyper-X is also considerably more efficient in terms of parameters and resources compared to methods that train separate adapters. Finally, Hyper-X consistently produces strong results in few-shot scenarios for new languages, showing the versatility of our approach beyond zero-shot transfer.

Some narratologists have argued that all stories derive from a limited set of archetypes. Specifically, Vonnegut (2005) claims in his Shapes of Stories lecture that if we graph the emotions in a story over time, the shape will be an instance of one of six basic story shapes. The work of Jockers (2015) and Reagan et al. (2016) purports to confirm this hypothesis empirically using automatic sentiment analysis (rather than manual annotations of story arcs) and algorithms to cluster story arcs into fundamental shapes. Later work has applied similar techniques to movies (Del Vecchio et al., 2019). This line of work has attracted criticism. Swafford (2015) argues that sentiment analysis needs to be validated on and adapted to narrative text. Enderle (2016) argues that the various methods to reduce story shapes to the putative six fundamental types are actually producing algorithmic artifacts, and that random sentiment arcs can also be clustered into six “fundamental” shapes. In this paper I will not attempt to find fundamental (or even universal) story shapes, but I will take the observed story shape for each narrative as is, without trying to cluster them into archetypes. My aim is to perform an empirical validation of how well basic sentiment analysis tools can reproduce a sentiment arc obtained through manual annotation based on narrative text. Rather than considering novels as narratives, I consider movies, since the annotation of movies, when done in real time, is less time consuming. In a previous abstract, I considered the task of predicting the annotated sentiment of individual sentences from movie scripts (van Cranenburgh, 2020), and concluded that sentiment analysis tools achieve comparable performance on narrative text as compared to reviews and social media text (pace Swafford 2015). In this abstract I consider the task of predicting the overall sentiment as annotated based on watching the movie. This task is more challenging since the connection between the narrative sentiment and the narrative text is potentially more distant.

Creating datasets for language phenomena to fill gaps in the language resource panorama of specific natural languages is not a trivial task. In this work, we explore the application of transferlearning as strategy to boost both the creation of language-specific datasets and systems. We use offensive language in Dutch tweets directed at Dutch politicians as a case study. In particular, we trained a multilingual model using the Political Speech Project (Bröckling et al. 2018) dataset to automatically annotate tweets in Dutch. The automatically annotated tweets have been used to further train a monolingual language model in Dutch (BERTje) adopting different strategies and combination of manually curated data. Our results show that: (i) transfer learning is an effective strategy to boost the creation of new datasets for specific language phenomena by reducing the annotation efforts; (ii) using a monolingual language model fine-tuned with automatically annotated data (i.e., silver data) is a competitive baseline against the zero-shot transfer of a multilingual model; and finally, (iii) less surprisingly, the addition of automatically annotated data to manually curated ones is a source of errors for the systems, degrading their performances.

We introduce the project MaCoCu: Massive collection and curation of monolingual and bilingual data: focus on under-resourced languages, funded by the Connecting Europe Facility, which is aimed at building monolingual and parallel corpora for under-resourced European languages. The approach followed consists of crawling large amounts of textual data from selected top-level domains of the Internet, and then applying a curation and enrichment pipeline. In addition to corpora, the project will release the free/open-source web crawling and curation software used.

We support the idea of applying cultural evolution theory to the study of storytelling, and fiction in particular. However, we suggest that a more plausible link between real and imaginary worlds is the feeling of “presence” we can experience in both of them: we feel present when we are able to correctly and intuitively enact our embodied predictions.

Current knowledge about the respiratory microbiome is mainly based on 16S ribosomal RNA gene sequencing. Newer sequencing approaches, such as metatranscriptomics, offer the technical ability to measure the viable microbiome response to environmental conditions such as smoking as well as to explore its functional role by investigating host-microbiome interactions. However, knowledge about its feasibility in respiratory microbiome research, especially in lung biopsies, is still very limited. RNA sequencing was performed in bronchial biopsies from clinically stable smokers (n=5) and ex-smokers (n=6) with COPD not using (inhaled) steroids. The Trinity assembler was used to assemble non-human reads in order to allow unbiased taxonomical and microbial transcriptional analyses. Subsequently, host-microbiome interactions were analyzed based on associations with host transcriptomic data. Ultra-low levels of microbial mass (0.009%) were identified in the RNA-seq data. Overall, no differences were identified in microbiome diversity or transcriptional profiles of microbial communities or individual microbes between COPD smokers and ex-smokers in the initial test dataset as well as a larger replication dataset. We identified an upregulated host gene set, related to the simultaneous presence of Bradyrhizobium, Roseomonas, Brevibacterium.spp., which were related to PERK-mediated unfolded protein response (UPR) and expression of the microRNA-155-5p. Our results show that metatranscriptomic profiling in bronchial biopsy samples from stable COPD patients yields ultra-low levels of microbial mass. Further, this study illustrates the potential of using transcriptional profiling of the host and microbiome to gain more insight into their interaction in the airways.