Language Resources and Evaluation最新文献

Due to the rise of user-generated content, social media is increasingly adopted as a channel to deliver customer service. Given the public character of online platforms, the automatic detection of emotions forms an important application in monitoring customer satisfaction and preventing negative word-of-mouth. This paper introduces EmoTwiCS, a corpus of 9489 Dutch customer service dialogues on Twitter that are annotated for emotion trajectories. In our business-oriented corpus, we view emotions as dynamic attributes of the customer that can change at each utterance of the conversation. The term ‘emotion trajectory’ refers therefore not only to the fine-grained emotions experienced by customers (annotated with 28 labels and valence-arousal-dominance scores), but also to the event happening prior to the conversation and the responses made by the human operator (both annotated with 8 categories). Inter-annotator agreement (IAA) scores on the resulting dataset are substantial and comparable with related research, underscoring its high quality. Given the interplay between the different layers of annotated information, we perform several in-depth analyses to investigate (i) static emotions in isolated tweets, (ii) dynamic emotions and their shifts in trajectory, and (iii) the role of causes and response strategies in emotion trajectories. We conclude by listing the advantages and limitations of our dataset, after which we give some suggestions on the different types of predictive modelling tasks and open research questions to which EmoTwiCS can be applied. The dataset is made publicly available at https://lt3.ugent.be/resources/emotwics.

Nowadays, social networks play a fundamental role in promoting and diffusing television and radio programs to different categories of audiences. So, political parties, influential groups and political activists have rapidly seized these new communication media to spread their ideas and give their sentiments concerning critical issues. In this context, Twitter, Facebook and YouTube have become very popular tools for sharing videos and communicating with users who interact with each other to discuss some problems, propose solutions and give viewpoints. This interaction on the social media sites yields to a huge amount of unstructured and noisy texts; hence the need for automated analysis techniques to classify sentiments conveyed in the users’ comments. In this work, we focus on opinions written in a less resourced Arabic language: Tunisian dialect (TD). In this work, we present a process for building a sentiment analyses model for comments written on Tunisian television broadcasts published in social media. These comments are written in a particular way with different spellings due to the fact that the Tunisian Dialect (TD) does not have an orthographic standard. For this we design crucial resources, namely sentiment lexicon and annotated corpus that we have used to investigate machine-learning and deep-learning models in order to identify the best sentiment analysis model for Tunisian Dialect.

Persian poetry has profoundly affected all periods of Persian literature and the literature of other countries as well. It is a fundamental vehicle for expressing Persian culture and political opinion. This paper presents a corpus of Persian literary text mainly focusing on poetry, covering the ninth to twenty-first century annotated for century and style, with additional partial annotation of rhetorical figures. Our resource is the largest and the most diverse corpus available in Persian literary text, with a particularly broad temporal scope. This allows us to conduct several computational experiments to analyze poetic styles, authors and time periods, as well as context shifts over time, for which we rely both on supervised models and on Persian poetry-specific heuristics. The corpus, the tools, and experiments described in this paper can be used not only for digital humanities studies of Persian literature but also for processing Persian texts in general, as well as in other broader cross-linguistic applications.

Learner corpora—datasets that reflect the language of non-native speakers—are instrumental for research of language learning and development, as well as for practical applications, mainly for teaching and education. Such corpora now exist for a plethora of native–foreign language pairs; but until recently, none of them reflected native Hebrew speakers, and very few reflected native Arabic speakers. We introduce a recently-released corpus of English essays authored by learners in Israel. The corpus consists of two sub-corpora, one of them of Arabic native speakers and the other consisting mainly of Hebrew native speakers. We report on the composition and curation of the datasets; specifically, we processed the data so that both sub-corpora are now uniformly represented, facilitating seamless research and computational processing of the data. We provide statistical information on the corpora and outline a few research projects that had already used them. This is the first and only learner corpus in Israel including two major native languages of people in the same educational system regarding the English syllabus. All the resources related to the corpus are freely available.

Most audio-visual (AV) emotion databases consist of clips that do not reflect real-life emotion processing (e.g., professional actors in bright studio-like environment), contain only spoken clips, and none have sung clips that express complex emotions. Here, we introduce a new AV database, the Reading Everyday Emotion Database (REED), which directly addresses those gaps. We recorded the faces of everyday adults with a diverse range of acting experience expressing 13 emotions—neutral, the six basic emotions (angry, disgusted, fearful, happy, sad, surprised), and six complex emotions (embarrassed, hopeful, jealous, proud, sarcastic, stressed)—in two auditory domains (spoken and sung) using everyday recording devices (e.g., laptops, mobile phones, etc.). The recordings were validated by an independent group of raters. We found that: intensity ratings of the recordings were positively associated with recognition accuracy; and the basic emotions, as well as the Neutral and Sarcastic emotions, were recognised more accurately than the other complex emotions. Emotion recognition accuracy also differed by utterance. Exploratory analysis revealed that recordings of those with drama experience were better recognised than those without. Overall, this database will benefit those who need AV clips with natural variations in both emotion expressions and recording environment.

In this paper, we discuss the development of a multilingual dataset annotated with a hierarchical, fine-grained tagset marking different types of aggression and the “context" in which they occur. The context, here, is defined by the conversational thread in which a specific comment occurs and also the “type” of discursive role that the comment is performing with respect to the previous comment(s). The dataset has been developed as part of the ComMA Project and consists of a total of 57,363 annotated comments, 1142 annotated memes, and around 70 h of annotated audio (extracted from videos) in four languages—Meitei, Bangla, Hindi, and Indian English. This data has been collected from various social media platforms such as YouTube, Facebook, Twitter, and Telegram. As is usual on social media websites, a large number of these comments are multilingual, and many are code-mixed with English. This paper gives a detailed description of the tagset developed during the course of this project and elaborates on the process of developing and using a multi-label, fine-grained tagset for marking comments with aggression and bias of various kinds, which includes gender bias, religious intolerance (called communal bias in the tagset), class/caste bias, and ethnic/racial bias. We define and discuss the tags that have been used for marking different discursive roles being performed through the comments, such as attack, defend, and so on. We also present a statistical analysis of the dataset as well as the results of our baseline experiments for developing an automatic aggression identification system using the dataset developed. Based on the results of the baseline experiments, we also argue that our dataset provides diverse and ‘hard’ sets of instances which makes it a good dataset for training and testing new techniques for aggressive and abusive language classification.

Automatic genre identification (AGI) is a text classification task focused on genres, i.e., text categories defined by the author’s purpose, common function of the text, and the text’s conventional form. Obtaining genre information has been shown to be beneficial for a wide range of disciplines, including linguistics, corpus linguistics, computational linguistics, natural language processing, information retrieval and information security. Consequently, in the past 20 years, numerous researchers have collected genre datasets with the aim to develop an efficient genre classifier. However, their approaches to the definition of genre schemata, data collection and manual annotation vary substantially, resulting in significantly different datasets. As most AGI experiments are dataset-dependent, a sufficient understanding of the differences between the available genre datasets is of great importance for the researchers venturing into this area. In this paper, we present a detailed overview of different approaches to each of the steps of the AGI task, from the definition of the genre concept and the genre schema, to the dataset collection and annotation methods, and, finally, to machine learning strategies. Special focus is dedicated to the description of the most relevant genre schemata and datasets, and details on the availability of all of the datasets are provided. In addition, the paper presents the recent advances in machine learning approaches to automatic genre identification, and concludes with proposing the directions towards developing a stable multilingual genre classifier.

This paper starts with an overview of corpora available for Brazilian Portuguese to subsequently focus mainly on the CoMET Project developed at the University of São Paulo. CoMET consists of three corpora: a comparable Portuguese-English technical corpus (CorTec), a Portuguese-English parallel (translation) corpus (CorTrad) and a multilingual learner corpus, (CoMAprend), all available for online queries with specific tools. CorTec offers over fifty corpora in a variety of domains, from Health Sciences to Olympic Games. CorTrad is divided into three parts: Popular Science, Technical-Scientific and Literary. Each one of CoMET’s corpora is presented in detail. Examples are also provided.