We introduce the first expert annotated corpus of Facebook comments for Hausa hate speech detection. The corpus titled HausaHate comprises 2,000 comments extracted from Western African Facebook pages and manually annotated by three Hausa native speakers, who are also NLP experts. Our corpus was annotated using two different layers. We first labeled each comment according to a binary classification: offensive versus non-offensive. Then, offensive comments were also labeled according to hate speech targets: race, gender and none. Lastly, a baseline model using fine-tuned LLM for Hausa hate speech detection is presented, highlighting the challenges of hate speech detection tasks for indigenous languages in Africa, as well as future advances.

Hate speech is a surely relevant problem in Brazil. Nevertheless, its regulation is not effective due to the difficulty to identify, quantify and classify offensive comments. Here, we introduce a novel system for offensive comment analysis in Brazilian Portuguese. The system titled “NoHateBrazil” recognizes explicit and implicit offensiveness in context at a fine-grained level. Specifically, we propose a framework for data collection, human annotation and machine learning models that were used to build the system. In addition, we assess the potential of our system to reflect stereotypical beliefs against marginalized groups by contrasting them with counter-stereotypes. As a result, a friendly web application was implemented, which besides presenting relevant performance, showed promising results towards mitigation of the risk of reinforcing social stereotypes. Lastly, new measures were proposed to improve the explainability of offensiveness classification and reliability of the model’s predictions.

Recent studies have shown that hate speech technologies may propagate social stereotypes against marginalized groups. Nevertheless, there has been a lack of realistic approaches to assess and mitigate biased technologies. In this paper, we introduce a new approach to analyze the potential of hate-speech classifiers to reflect social stereotypes through the investigation of stereotypical beliefs by contrasting them with counter-stereotypes. We empirically measure the distribution of stereotypical beliefs by analyzing the distinctive classification of tuples containing stereotypes versus counter-stereotypes in machine learning models and datasets. Experiment results show that hate speech classifiers attribute unreal or negligent offensiveness to social identity groups by reflecting and reinforcing stereotypical beliefs regarding minorities. Furthermore, we also found that models that embed expert and context information from offensiveness markers present promising results to mitigate social stereotype bias towards socially responsible hate speech detection.

Automated news credibility and fact-checking at scale require accurate prediction of news factuality and media bias. This paper introduces a large sentence-level dataset, titled “FactNews”, composed of 6,191 sentences expertly annotated according to factuality and media bias definitions proposed by AllSides. We use FactNews to assess the overall reliability of news sources by formulating two text classification problems for predicting sentence-level factuality of news reporting and bias of media outlets. Our experiments demonstrate that biased sentences present a higher number of words compared to factual sentences, besides having a predominance of emotions. Hence, the fine-grained analysis of subjectivity and impartiality of news articles showed promising results for predicting the reliability of entire media outlets. Finally, due to the severity of fake news and political polarization in Brazil, and the lack of research for Portuguese, both dataset and baseline were proposed for Brazilian Portuguese.

Over the years, the review helpfulness prediction task has been the subject of several works, but remains being a challenging issue in Natural Language Processing, as results vary a lot depending on the domain, on the adopted features and on the chosen classification strategy. This paper attempts to evaluate the impact of content features and classification methods for two different domains. In particular, we run our experiments for a low resource language – Portuguese –, trying to establish a benchmark for this language. We show that simple features and classical classification methods are powerful for the task of helpfulness prediction, but are largely outperformed by a convolutional neural network-based solution.

This work presents a two-stage approach for tackling low-resource AMR-to-text generation for Brazilian Portuguese. Our approach consists of (1) generating a masked surface realization in which some tokens are masked according to its Part-of-Speech class and (2) infilling the masked tokens according to the AMR graph and the previous masked surface realization. Results show a slight improvement over the baseline, mainly in BLEU (1.63) and METEOR (0.02) scores. Moreover, we evaluate the pipeline components separately, showing that the bottleneck of the pipeline is the masked surface realization. Finally, the human evaluation suggests that models still suffer from hallucinations, and some strategies to deal with the problems found are proposed.

Most information is passed on in the form of language. Therefore, research on how people use language to inform and misinform, and how this knowledge may be automatically extracted from large amounts of text is surely relevant. This survey provides first-hand experiences and a comprehensive review of rhetorical-level structure analysis for online deception detection. We systematically analyze how discourse structure, aligned or not with other approaches, is applied to automatic fake news and fake reviews detection on the web and social media. Moreover, we categorize discourse-tagged corpora along with results, hence offering a summary and accessible introductions to new researchers.

This paper presents PortiLexicon-UD, a large and freely available lexicon for Portuguese delivering morphosyntactic information according to the Universal Dependencies model. This lexical resource includes part of speech tags, lemmas, and morphological information for words, with 1,221,218 entries (considering word duplication due to different combination of PoS tag, lemma, and morphological features). We report the lexicon creation process, its computational data structure, and its evaluation over an annotated corpus, showing that it has a high language coverage and good quality data.

Due to the severity of the social media offensive and hateful comments in Brazil, and the lack of research in Portuguese, this paper provides the first large-scale expert annotated corpus of Brazilian Instagram comments for hate speech and offensive language detection. The HateBR corpus was collected from the comment section of Brazilian politicians’ accounts on Instagram and manually annotated by specialists, reaching a high inter-annotator agreement. The corpus consists of 7,000 documents annotated according to three different layers: a binary classification (offensive versus non-offensive comments), offensiveness-level classification (highly, moderately, and slightly offensive), and nine hate speech groups (xenophobia, racism, homophobia, sexism, religious intolerance, partyism, apology for the dictatorship, antisemitism, and fatphobia). We also implemented baseline experiments for offensive language and hate speech detection and compared them with a literature baseline. Results show that the baseline experiments on our corpus outperform the current state-of-the-art for the Portuguese language.

The amount of information available online can be overwhelming for users to digest, specially when dealing with other users’ comments when making a decision about buying a product or service. In this context, opinion summarization systems are of great value, extracting important information from the texts and presenting them to the user in a more understandable manner. It is also known that the usage of semantic representations can benefit the quality of the generated summaries. This paper aims at developing opinion summarization methods based on Abstract Meaning Representation of texts in the Brazilian Portuguese language. Four different methods have been investigated, alongside some literature approaches. The results show that a Machine Learning-based method produced summaries of higher quality, outperforming other literature techniques on manually constructed semantic graphs. We also show that using parsed graphs over manually annotated ones harmed the output. Finally, an analysis of how important different types of information are for the summarization process suggests that using Sentiment Analysis features did not improve summary quality.

Since a lexicon-based approach is more elegant scientifically, explaining the solution components and being easier to generalize to other applications, this paper provides a new approach for offensive language and hate speech detection on social media, which embodies a lexicon of implicit and explicit offensive and swearing expressions annotated with contextual information. Due to the severity of the social media abusive comments in Brazil, and the lack of research in Portuguese, Brazilian Portuguese is the language used to validate the models. Nevertheless, our method may be applied to any other language. The conducted experiments show the effectiveness of the proposed approach, outperforming the current baseline methods for the Portuguese language.

Statements that are intentionally misstated (or manipulated) are of considerable interest to researchers, government, security, and financial systems. According to deception literature, there are reliable cues for detecting deception and the belief that liars give off cues that may indicate their deception is near-universal. Therefore, given that deceiving actions require advanced cognitive development that honesty simply does not require, as well as people’s cognitive mechanisms have promising guidance for deception detection, in this Ph.D. ongoing research, we propose to examine discourse structure patterns in multilingual deceptive news corpora using the Rhetorical Structure Theory framework. Considering that our work is the first to exploit multilingual discourse-aware strategies for fake news detection, the research community currently lacks multilingual deceptive annotated corpora. Accordingly, this paper describes the current progress in this thesis, including (i) the construction of the first multilingual deceptive corpus, which was annotated by specialists according to the Rhetorical Structure Theory framework, and (ii) the introduction of two new proposed rhetorical relations: INTERJECTION and IMPERATIVE, which we assume to be relevant for the fake news detection task.

This paper describes the submission by the NILC Computational Linguistics research group of the University of S ̃ao Paulo/Brazil to the English Track 2 (closed sub-track) at the Surface Realisation Shared Task 2020. The success of the current pre-trained models like BERT or GPT-2 in several tasks is well-known, however, this is not the case for data-to-text generation tasks and just recently some initiatives focused on it. This way, we explore how a pre-trained model (GPT-2) performs on the UD-to-text generation task. In general, the achieved results were poor, but there are some interesting ideas to explore. Among the learned lessons we may note that it is necessary to study strategies to represent UD inputs and to introduce structural knowledge into these pre-trained models.

The proliferation of fake news is a current issue that influences a number of important areas of society, such as politics, economy and health. In the Natural Language Processing area, recent initiatives tried to detect fake news in different ways, ranging from language-based approaches to content-based verification. In such approaches, the choice of the features for the classification of fake and true news is one of the most important parts of the process. This paper presents a study on the impact of readability features to detect fake news for the Brazilian Portuguese language. The results show that such features are relevant to the task (achieving, alone, up to 92% classification accuracy) and may improve previous classification results.

Abstract Meaning Representation (AMR) is a graph-based semantic formalism where the nodes are concepts and edges are relations among them. Most of AMR parsing methods require alignment between the nodes of the graph and the words of the sentence. However, this alignment is not provided by manual annotations and available automatic aligners focus only on the English language, not performing well for other languages. Aiming to fulfill this gap, we developed an alignment method for the Portuguese language based on a more semantically matched word-concept pair. We performed both intrinsic and extrinsic evaluations and showed that our alignment approach outperforms the alignment strategies developed for English, improving AMR parsers, and achieving competitive results with a parser designed for the Portuguese language.

Abstract Meaning Representation (AMR) is a recent and prominent semantic representation with good acceptance and several applications in the Natural Language Processing area. For English, there is a large annotated corpus (with approximately 39K sentences) that supports the research with the representation. However, to the best of our knowledge, there is only one restricted corpus for Portuguese, which contains 1,527 sentences. In this context, this paper presents an effort to build a general purpose AMR-annotated corpus for Brazilian Portuguese by translating and adapting AMR English guidelines. Our results show that such approach is feasible, but there are some challenging phenomena to solve. More than this, efforts are necessary to increase the coverage of the corresponding lexical resource that supports the annotation.

This paper presents an exploratory study that aims to evaluate the usefulness of back-translation in Natural Language Generation (NLG) from semantic representations for non-English languages. Specifically, Abstract Meaning Representation and Brazilian Portuguese (BP) are chosen as semantic representation and language, respectively. Two methods (focused on Statistical and Neural Machine Translation) are evaluated on two datasets (one automatically generated and another one human-generated) to compare the performance in a real context. Also, several cuts according to quality measures are performed to evaluate the importance (or not) of the data quality in NLG. Results show that there are still many improvements to be made but this is a promising approach.

This paper presents a more recent literature review on Natural Language Generation. In particular, we highlight the efforts for Brazilian Portuguese in order to show the available resources and the existent approaches for this language. We also focus on the approaches for generation from semantic representations (emphasizing the Abstract Meaning Representation formalism) as well as their advantages and limitations, including possible future directions.

This paper describes the submission by the NILC Computational Linguistics research group of the University of São Paulo/Brazil to the Track 1 of the Surface Realization Shared Task (SRST Track 1). We present a neural-based method that works at the syntactic level to order the words (which we refer by NILC-SWORNEMO, standing for “Syntax-based Word ORdering using NEural MOdels”). Additionally, we apply a bottom-up approach to build the sentence and, using language-specific lexicons, we produce the proper word form of each lemma in the sentence. The results obtained by our method outperformed the average of the results for English, Portuguese and Spanish in the track.

Web 2.0 has allowed a never imagined communication boom. With the widespread use of computational and mobile devices, anyone, in practically any language, may post comments in the web. As such, formal language is not necessarily used. In fact, in these communicative situations, language is marked by the absence of more complex syntactic structures and the presence of internet slang, with missing diacritics, repetitions of vowels, and the use of chat-speak style abbreviations, emoticons and colloquial expressions. Such language use poses severe new challenges for Natural Language Processing (NLP) tools and applications, which, so far, have focused on well-written texts. In this work, we report the construction of a large web corpus of product reviews in Brazilian Portuguese and the analysis of its lexical phenomena, which support the development of a lexical normalization tool for, in future work, subsidizing the use of standard NLP products for web opinion mining and summarization purposes.