This paper presents our results for the Arabic Financial NLP (AraFinNLP) shared task at the Second Arabic Natural Language Processing Conference (ArabicNLP 2024). We participated in the first sub-task, Multi-dialect Intent Detection, which focused on cross-dialect intent detection in the banking domain. Our approach involved fine-tuning an encoder-only T5 model, generating synthetic data, and model ensembling. Additionally, we conducted an in-depth analysis of the dataset, addressing annotation errors and problematic translations. Our model was ranked third in the shared task, achieving a F1-score of 0.871.
We present EDEN, the first Norwegian dataset annotated with event information at the sentence level, adapting the widely used ACE event schema to Norwegian. The paper describes the manual annotation of Norwegian text as well as transcribed speech in the news domain, together with inter-annotator agreement and discussions of relevant dataset statistics. We also present preliminary modeling results using a graph-based event parser. The resulting dataset will be freely available for download and use.
We investigate in this paper how distributions of occupations with respect to gender is reflected in pre-trained language models. Such distributions are not always aligned to normative ideals, nor do they necessarily reflect a descriptive assessment of reality. In this paper, we introduce an approach for measuring to what degree pre-trained language models are aligned to normative and descriptive occupational distributions. To this end, we use official demographic information about gender–occupation distributions provided by the national statistics agencies of France, Norway, United Kingdom, and the United States. We manually generate template-based sentences combining gendered pronouns and nouns with occupations, and subsequently probe a selection of ten language models covering the English, French, and Norwegian languages. The scoring system we introduce in this work is language independent, and can be used on any combination of template-based sentences, occupations, and languages. The approach could also be extended to other dimensions of national census data and other demographic variables.
We propose a graph-based event extraction framework JSEEGraph that approaches the task of event extraction as general graph parsing in the tradition of Meaning Representation Parsing. It explicitly encodes entities and events in a single semantic graph, and further has the flexibility to encode a wider range of additional IE relations and jointly infer individual tasks. JSEEGraph performs in an end-to-end manner via general graph parsing: (1) instead of flat sequence labelling, nested structures between entities/triggers are efficiently encoded as separate nodes in the graph, allowing for nested and overlapping entities and triggers; (2) both entities, relations, and events can be encoded in the same graph, where entities and event triggers are represented as nodes and entity relations and event arguments are constructed via edges; (3) joint inference avoids error propagation and enhances the interpolation of different IE tasks. We experiment on two benchmark datasets of varying structural complexities; ACE05 and Rich ERE, covering three languages: English, Chinese, and Spanish. Experimental results show that JSEEGraph can handle nested event structures, that it is beneficial to solve different IE tasks jointly, and that event argument extraction in particular benefits from entity extraction. Our code and models are released as open-source.
This paper provides a systematic analysis and comparison of the performance of state-of-the-art models on the task of fine-grained Arabic dialect identification using the MADAR parallel corpus. We test approaches based on pre-trained transformer language models in addition to Naive Bayes models with a rich set of various features. Through a comprehensive data- and error analysis, we provide valuable insights into the strengths and weaknesses of both approaches. We discuss which dialects are more challenging to differentiate, and identify potential sources of errors. Our analysis reveals an important problem with identical sentences across dialect classes in the test set of the MADAR-26 corpus, which may confuse any classifier. We also show that none of the tested approaches captures the subtle distinctions between closely related dialects.
We investigate to what extent pre-trained language models can be used for automated claim detection for fact-checking in a low resource setting. We explore this idea by fine-tuning four Norwegian pre-trained language models to perform the binary classification task of determining if a claim should be discarded or upheld to be further processed by human fact-checkers. We conduct a set of experiments to compare the performance of the language models, and provide a simple baseline model using SVM with tf-idf features. Since we are focusing on claim detection, the recall score for the upheld class is to be emphasized over other performance measures. Our experiments indicate that the language models are superior to the baseline system in terms of F1, while the baseline model results in the highest precision. However, the two Norwegian models, NorBERT2 and NB-BERT_large, give respectively superior F1 and recall values. We argue that large language models could be successfully employed to solve the automated claim detection problem. The choice of the model depends on the desired end-goal. Moreover, our error analysis shows that language models are generally less sensitive to the changes in claim length and source than the SVM model.
Dialectal variation is present in many human languages and is attracting a growing interest in NLP. Most previous work concentrated on either (1) classifying dialectal varieties at the document or sentence level or (2) performing standard NLP tasks on dialectal data. In this paper, we propose the novel task of token-level dialectal feature prediction. We present a set of fine-grained annotation guidelines for Norwegian dialects, expand a corpus of dialectal tweets, and manually annotate them using the introduced guidelines. Furthermore, to evaluate the learnability of our task, we conduct labeling experiments using a collection of baselines, weakly supervised and supervised sequence labeling models. The obtained results show that, despite the difficulty of the task and the scarcity of training data, many dialectal features can be predicted with reasonably high accuracy.
We present NorBench: a streamlined suite of NLP tasks and probes for evaluating Norwegian language models (LMs) on standardized data splits and evaluation metrics. We also introduce a range of new Norwegian language models (both encoder and encoder-decoder based). Finally, we compare and analyze their performance, along with other existing LMs, across the different benchmark tests of NorBench.
In this paper we explore how a demographic distribution of occupations, along gender dimensions, is reflected in pre-trained language models. We give a descriptive assessment of the distribution of occupations, and investigate to what extent these are reflected in four Norwegian and two multilingual models. To this end, we introduce a set of simple bias probes, and perform five different tasks combining gendered pronouns, first names, and a set of occupations from the Norwegian statistics bureau. We show that language specific models obtain more accurate results, and are much closer to the real-world distribution of clearly gendered occupations. However, we see that none of the models have correct representations of the occupations that are demographically balanced between genders. We also discuss the importance of the training data on which the models were trained on, and argue that template-based bias probes can sometimes be fragile, and a simple alteration in a template can change a model’s behavior.
Norwegian Twitter data poses an interesting challenge for Natural Language Processing (NLP) tasks. These texts are difficult for models trained on standardized text in one of the two Norwegian written forms (Bokmål and Nynorsk), as they contain both the typical variation of social media text, as well as a large amount of dialectal variety. In this paper we present a novel Norwegian Twitter dataset annotated with POS-tags. We show that models trained on Universal Dependency (UD) data perform worse when evaluated against this dataset, and that models trained on Bokmål generally perform better than those trained on Nynorsk. We also see that performance on dialectal tweets is comparable to the written standards for some models. Finally we perform a detailed analysis of the errors that models commonly make on this data.
Scandinavian countries are perceived as role-models when it comes to gender equality. With the advent of pre-trained language models and their widespread usage, we investigate to what extent gender-based harmful and toxic content exists in selected Scandinavian language models. We examine nine models, covering Danish, Swedish, and Norwegian, by manually creating template-based sentences and probing the models for completion. We evaluate the completions using two methods for measuring harmful and toxic completions and provide a thorough analysis of the results. We show that Scandinavian pre-trained language models contain harmful and gender-based stereotypes with similar values across all languages. This finding goes against the general expectations related to gender equality in Scandinavian countries and shows the possible problematic outcomes of using such models in real-world settings. Warning: Some of the examples provided in this paper can be upsetting and offensive.
We describe NorDiaChange: the first diachronic semantic change dataset for Norwegian. NorDiaChange comprises two novel subsets, covering about 80 Norwegian nouns manually annotated with graded semantic change over time. Both datasets follow the same annotation procedure and can be used interchangeably as train and test splits for each other. NorDiaChange covers the time periods related to pre- and post-war events, oil and gas discovery in Norway, and technological developments. The annotation was done using the DURel framework and two large historical Norwegian corpora. NorDiaChange is published in full under a permissive licence, complete with raw annotation data and inferred diachronic word usage graphs (DWUGs).
This paper introduces a first step towards creating the NERDz dataset. A manually annotated dataset of named entities for the Algerian vernacular dialect. The annotations are built on top of a recent extension to the Algerian NArabizi Treebank, comprizing NArabizi sentences with manual transliterations into Arabic and code-switched scripts. NERDz is therefore not only the first dataset of named entities for Algerian, but it also comprises parallel entities written in Latin, Arabic, and code-switched scripts. We present a detailed overview of our annotations, inter-annotator agreement measures, and define two preliminary baselines using a neural sequence labeling approach and an Algerian BERT model. We also make the annotation guidelines and the annotations available for future work
We investigate in this paper how correlations between occupations and gendered-pronouns can be affected and changed by adding negation in bias probes, or changing the grammatical tense of the verbs in the probes. We use a set of simple bias probes in Norwegian and English, and perform 16 different probing analysis, using four Norwegian and four English pre-trained language models. We show that adding negation to probes does not have a considerable effect on the correlations between gendered-pronouns and occupations, supporting other works on negation in language models. We also show that altering the grammatical tense of verbs in bias probes do have some interesting effects on models’ behaviours and correlations. We argue that we should take grammatical tense into account when choosing bias probes, and aggregating results across tenses might be a better representation of the existing correlations.
Event extraction involves the detection and extraction of both the event triggers and the corresponding arguments. Existing systems often decompose event extraction into multiple subtasks, without considering their possible interactions. In this paper, we propose EventGraph, a joint framework for event extraction, which encodes events as graphs. We represent event triggers and arguments as nodes in a semantic graph. Event extraction therefore becomes a graph parsing problem, which provides the following advantages: 1) performing event detection and argument extraction jointly; 2) detecting and extracting multiple events from a piece of text; 3) capturing the complicated interaction between event arguments and triggers. Experimental results on ACE2005 show that our model is competitive to state-of-the-art systems and has substantially improved the results on argument extraction. Additionally, we create two new datasets from ACE2005 where we keep the entire text spans for event arguments, instead of just the head word(s). Our code and models will be released as open-source.
This paper presents our submission to the 2022 edition of the CASE 2021 shared task 1, subtask 4. The EventGraph system adapts an end-to-end, graph-based semantic parser to the task of Protest Event Extraction and more specifically subtask 4 on event trigger and argument extraction. We experiment with various graphs, encoding the events as either “labeled-edge” or “node-centric” graphs. We show that the “node-centric” approach yields best results overall, performing well across the three languages of the task, namely English, Spanish, and Portuguese. EventGraph is ranked 3rd for English and Portuguese, and 4th for Spanish.
Norway has a large amount of dialectal variation, as well as a general tolerance to its use in the public sphere. There are, however, few available resources to study this variation and its change over time and in more informal areas, on social media. In this paper, we propose a first step to creating a corpus of dialectal variation of written Norwegian. We collect a small corpus of tweets and manually annotate them as Bokmål, Nynorsk, any dialect, or a mix. We further perform preliminary experiments with state-of-the-art models, as well as an analysis of the data to expand this corpus in the future. Finally, we make the annotations available for future work.
In this work we explore the effect of incorporating demographic metadata in a text classifier trained on top of a pre-trained transformer language model. More specifically, we add information about the gender of critics and book authors when classifying the polarity of book reviews, and the polarity of the reviews when classifying the genders of authors and critics. We use an existing data set of Norwegian book reviews with ratings by professional critics, which has also been augmented with gender information, and train a document-level sentiment classifier on top of a recently released Norwegian BERT-model. We show that gender-informed models obtain substantially higher accuracy, and that polarity-informed models obtain higher accuracy when classifying the genders of book authors. For this particular data set, we take this result as a confirmation of the gender bias in the underlying label distribution, but in other settings we believe a similar approach can be used for mitigating bias in the model.
Gender bias in models and datasets is widely studied in NLP. The focus has usually been on analysing how females and males express themselves, or how females and males are described. However, a less studied aspect is the combination of these two perspectives, how female and male describe the same or opposite gender. In this paper, we present a new gender annotated sentiment dataset of critics reviewing the works of female and male authors. We investigate if this newly annotated dataset contains differences in how the works of male and female authors are critiqued, in particular in terms of positive and negative sentiment. We also explore the differences in how this is done by male and female critics. We show that there are differences in how critics assess the works of authors of the same or opposite gender. For example, male critics rate crime novels written by females, and romantic and sentimental works written by males, more negatively.
We present in this paper our work on Algerian language, an under-resourced North African colloquial Arabic variety, for which we built a comparably large corpus of more than 36,000 code-switched user-generated comments annotated for sentiments. We opted for this data domain because Algerian is a colloquial language with no existing freely available corpora. Moreover, we compiled sentiment lexicons of positive and negative unigrams and bigrams reflecting the code-switches present in the language. We compare the performance of four models on the task of identifying sentiments, and the results indicate that a CNN model trained end-to-end fits better our unedited code-switched and unbalanced data across the predefined sentiment classes. Additionally, injecting the lexicons as background knowledge to the model boosts its performance on the minority class with a gain of 10.54 points on the F-score. The results of our experiments can be used as a baseline for future research for Algerian sentiment analysis.
Named Entity Recognition (NER) performance often degrades rapidly when applied to target domains that differ from the texts observed during training. When in-domain labelled data is available, transfer learning techniques can be used to adapt existing NER models to the target domain. But what should one do when there is no hand-labelled data for the target domain? This paper presents a simple but powerful approach to learn NER models in the absence of labelled data through weak supervision. The approach relies on a broad spectrum of labelling functions to automatically annotate texts from the target domain. These annotations are then merged together using a hidden Markov model which captures the varying accuracies and confusions of the labelling functions. A sequence labelling model can finally be trained on the basis of this unified annotation. We evaluate the approach on two English datasets (CoNLL 2003 and news articles from Reuters and Bloomberg) and demonstrate an improvement of about 7 percentage points in entity-level F1 scores compared to an out-of-domain neural NER model.
This paper presents our results for the Nuanced Arabic Dialect Identification (NADI) shared task of the Fifth Workshop for Arabic Natural Language Processing (WANLP 2020). We participated in the first sub-task for country-level Arabic dialect identification covering 21 Arab countries. Our contribution is based on a stacking classifier using Multinomial Naive Bayes, Linear SVC, and Logistic Regression classifiers as estimators; followed by a Logistic Regression as final estimator. Despite the fact that the results on the test set were low, with a macro F1 of 17.71, we were able to show that a simple approach can achieve comparable results to more sophisticated solutions. Moreover, the insights of our error analysis, and of the corpus content in general, can be used to develop and improve future systems.
We measure the intensity of diachronic semantic shifts in adjectives in English, Norwegian and Russian across 5 decades. This is done in order to test the hypothesis that evaluative adjectives are more prone to temporal semantic change. To this end, 6 different methods of quantifying semantic change are used. Frequency-controlled experimental results show that, depending on the particular method, evaluative adjectives either do not differ from other types of adjectives in terms of semantic change or appear to actually be less prone to shifting (particularly, to ‘jitter’-type shifting). Thus, in spite of many well-known examples of semantically changing evaluative adjectives (like ‘terrific’ or ‘incredible’), it seems that such cases are not specific to this particular type of words.
This paper explores the use of multi-task learning (MTL) for incorporating external knowledge in neural models. Specifically, we show how MTL can enable a BiLSTM sentiment classifier to incorporate information from sentiment lexicons. Our MTL set-up is shown to improve model performance (compared to a single-task set-up) on both English and Norwegian sentence-level sentiment datasets. The paper also introduces a new sentiment lexicon for Norwegian.
Automatically identifying persons in a particular role within a large corpus can be a difficult task, especially if you don’t know who you are actually looking for. Resources compiling names of persons can be available, but no exhaustive lists exist. However, such lists usually contain known names that are “visible” in the national public sphere, and tend to ignore the marginal and international ones. In this article we propose a method for automatically generating suggestions of names found in a corpus of Norwegian news articles, and which “naturally” belong to a given initial list of members, and that were not known (compiled in a list) beforehand. The approach is based, in part, on the assumption that surface level syntactic features reveal parts of the underlying semantic content and can help uncover the structure of the language.