The paper aims to detect and mitigate LGBTQIA+ bias in large language models (LLMs). As the usage of LLMs quickly increases, so does the significance of the harms they may cause due to bias. The research field of bias in LLMs has seen massive growth, but few attempts have been made to detect or mitigate other biases than gender bias, and most focus has been on English LLMs. This work shows experimentally that LLMs may cause representational harms towards LGBTQIA+ individuals when evaluated on sentence completion tasks and on a benchmark dataset constructed from stereotypes reported by the queer community of Norway, collected through a survey in order to directly involve the affected community. Furthermore, Norwegian training corpora are probed for queer bias, revealing strong associations between queer terms and anti-queer slurs, as well as words related to pedophilia. Finally, a fine-tuning-based debiasing method is applied to two Norwegian LLMs. This method does not consistently reduce bias, but shows that queer bias can be altered, laying the foundation for future debiasing approaches. By shedding light on the severe discrimination that can occur through the usage of LLMs, this paper contributes to the ongoing fight for equal rights for the LGBTQIA+ community.

The paper introduces a cross-lingual speaker identification system for Indian languages, utilising a Long Short-Term Memory dense neural network (LSTM-DNN). The system was trained on audio recordings in English and evaluated on data from Hindi, Kannada, Malayalam, Tamil, and Telugu, with a view to how factors such as phonetic similarity and native accent affect performance. The model was fed with MFCC (mel-frequency cepstral coefficient) features extracted from the audio file. For comparison, the corresponding mel-spectrogram images were also used as input to a ResNet-50 model, while the raw audio was used to train a Siamese network. The LSTM-DNN model outperformed the other two models as well as two more traditional baseline speaker identification models, showing that deep learning models are superior to probabilistic models for capturing low-level speech features and learning speaker characteristics.

Distinguishing hate speech from non-hate offensive language is challenging, as hate speech not always includes offensive slurs and offensive language not always express hate. Here, four deep learners based on the Bidirectional Encoder Representations from Transformers (BERT), with either general or domain-specific language models, were tested against two datasets containing tweets labelled as either ‘Hateful’, ‘Normal’ or ‘Offensive’. The results indicate that the attention-based models profoundly confuse hate speech with offensive and normal language. However, the pre-trained models outperform state-of-the-art results in terms of accurately predicting the hateful instances.

The paper investigates how well poetry can be generated to contain a specific sentiment, and whether readers of the poetry experience the intended sentiment. The poetry generator consists of a bi-directional Long Short-Term Memory (LSTM) model, combined with rhyme pair generation, rule-based word prediction methods, and tree search for extending generation possibilities. The LSTM network was trained on a set of English poetry written and published by users on a public website. Human judges evaluated poems generated by the system, both with a positive and negative sentiment. The results indicate that while there are some weaknesses in the system compared to other state-of-the-art solutions, it is fully capable of generating poetry with an inherent sentiment that is perceived by readers.

The paper explores how an attention-based approach can increase performance on the task of native-language identification (NLI), i.e., to identify an author’s first language given information expressed in a second language. Previously, Support Vector Machines have consistently outperformed deep learning-based methods on the TOEFL11 data set, the de facto standard for evaluating NLI systems. The attention-based system BERT (Bidirectional Encoder Representations from Transformers) was first tested in isolation on the TOEFL11 data set, then used in a meta-classifier stack in combination with traditional techniques to produce an accuracy of 0.853. However, more labelled NLI data is now available, so BERT was also trained on the much larger Reddit-L2 data set, containing 50 times as many examples as previously used for English NLI, giving an accuracy of 0.902 on the Reddit-L2 in-domain test scenario, improving the state-of-the-art by 21.2 percentage points.

Information on social media comprises of various modalities such as textual, visual and audio. NLP and Computer Vision communities often leverage only one prominent modality in isolation to study social media. However, computational processing of Internet memes needs a hybrid approach. The growing ubiquity of Internet memes on social media platforms such as Facebook, Instagram, and Twitter further suggests that we can not ignore such multimodal content anymore. To the best of our knowledge, there is not much attention towards meme emotion analysis. The objective of this proposal is to bring the attention of the research community towards the automatic processing of Internet memes. The task Memotion analysis released approx 10K annotated memes- with human annotated labels namely sentiment(positive, negative, neutral), type of emotion(sarcastic,funny,offensive, motivation) and their corresponding intensity. The challenge consisted of three subtasks: sentiment (positive, negative, and neutral) analysis of memes,overall emotion (humor, sarcasm, offensive, and motivational) classification of memes, and classifying intensity of meme emotion. The best performances achieved were F1 (macro average) scores of 0.35, 0.51 and 0.32, respectively for each of the three subtasks.

In this paper, we present the results of the SemEval-2020 Task 9 on Sentiment Analysis of Code-Mixed Tweets (SentiMix 2020). We also release and describe our Hinglish (Hindi-English)and Spanglish (Spanish-English) corpora annotated with word-level language identification and sentence-level sentiment labels. These corpora are comprised of 20K and 19K examples, respectively. The sentiment labels are - Positive, Negative, and Neutral. SentiMix attracted 89 submissions in total including 61 teams that participated in the Hinglish contest and 28 submitted systems to the Spanglish competition. The best performance achieved was 75.0% F1 score for Hinglish and 80.6% F1 for Spanglish. We observe that BERT-like models and ensemble methods are the most common and successful approaches among the participants.

Work on Abusive Language Detection has tackled a wide range of subtasks and domains. As a result of this, there exists a great deal of redundancy and non-generalisability between datasets. Through experiments on cross-dataset training and testing, the paper reveals that the preconceived notion of including more non-abusive samples in a dataset (to emulate reality) may have a detrimental effect on the generalisability of a model trained on that data. Hence a hierarchical annotation model is utilised here to reveal redundancies in existing datasets and to help reduce redundancy in future efforts.

The paper describes the systems submitted to OffensEval (SemEval 2019, Task 6) on ‘Identifying and Categorizing Offensive Language in Social Media’ by the ‘NIT_Agartala_NLP_Team’. A Twitter annotated dataset of 13,240 English tweets was provided by the task organizers to train the individual models, with the best results obtained using an ensemble model composed of six different classifiers. The ensemble model produced macro-averaged F1-scores of 0.7434, 0.7078 and 0.4853 on Subtasks A, B, and C, respectively. The paper highlights the overall low predictive nature of various linguistic features and surface level count features, as well as the limitations of a traditional machine learning approach when compared to a Deep Learning counterpart.

Hate speech detectors must be applicable across a multitude of services and platforms, and there is hence a need for detection approaches that do not depend on any information specific to a given platform. For instance, the information stored about the text’s author may differ between services, and so using such data would reduce a system’s general applicability. The paper thus focuses on using exclusively text-based input in the detection, in an optimised architecture combining Convolutional Neural Networks and Long Short-Term Memory-networks. The hate speech detector merges two strands with character n-grams and word embeddings to produce the final classification, and is shown to outperform comparable previous approaches.

The paper presents NTNU’s contribution to SemEval-2018 Task 7 on relation identification and classification. The class weights and parameters of five alternative supervised classifiers were optimized through grid search and cross-validation. The outputs of the classifiers were combined through voting for the final prediction. A wide variety of features were explored, with the most informative identified by feature selection. The best setting achieved F1 scores of 47.4% and 66.0% in the relation classification subtasks 1.1 and 1.2. For relation identification and classification in subtask 2, it achieved F1 scores of 33.9% and 17.0%,

Cybersecurity risks such as malware threaten the personal safety of users, but to identify malware text is a major challenge. The paper proposes a supervised learning approach to identifying malware sentences given a document (subTask1 of SemEval 2018, Task 8), as well as to classifying malware tokens in the sentences (subTask2). The approach achieved good results, ranking second of twelve participants for both subtasks, with F-scores of 57% for subTask1 and 28% for subTask2.

Named Entity Recognition is an important information extraction task that identifies proper names in unstructured texts and classifies them into some pre-defined categories. Identification of named entities in code-mixed social media texts is a more difficult and challenging task as the contexts are short, ambiguous and often noisy. This work proposes a Conditional Random Fields based named entity recognition system to identify proper names in code-switched data and classify them into nine categories. The system ranked fifth among nine participant systems and achieved a 59.25% F1-score.

The paper investigates the potential effects user features have on hate speech classification. A quantitative analysis of Twitter data was conducted to better understand user characteristics, but no correlations were found between hateful text and the characteristics of the users who had posted it. However, experiments with a hate speech classifier based on datasets from three different languages showed that combining certain user features with textual features gave slight improvements of classification performance. While the incorporation of user features resulted in varying impact on performance for the different datasets used, user network-related features provided the most consistent improvements.

The paper proposes the Ternary Sentiment Embedding Model, a new model for creating sentiment embeddings based on the Hybrid Ranking Model of Tang et al. (2016), but trained on ternary-labeled data instead of binary-labeled, utilizing sentiment embeddings from datasets made with different distant supervision methods. The model is used as part of a complete Twitter Sentiment Analysis system and empirically compared to existing systems, showing that it outperforms Hybrid Ranking and that the quality of the distant-supervised dataset has a great impact on the quality of the produced sentiment embeddings.

To find out how users’ social media behaviour and language are related to their ethical practices, the paper investigates applying Schwartz’ psycholinguistic model of societal sentiment to social media text. The analysis is based on corpora collected from user essays as well as social media (Facebook and Twitter). Several experiments were carried out on the corpora to classify the ethical values of users, incorporating Linguistic Inquiry Word Count analysis, n-grams, topic models, psycholinguistic lexica, speech-acts, and non-linguistic information, while applying a range of machine learners (Support Vector Machines, Logistic Regression, and Random Forests) to identify the best linguistic and non-linguistic features for automatic classification of values and ethics.

The paper introduces a deep learning-based Twitter hate-speech text classification system. The classifier assigns each tweet to one of four predefined categories: racism, sexism, both (racism and sexism) and non-hate-speech. Four Convolutional Neural Network models were trained on resp. character 4-grams, word vectors based on semantic information built using word2vec, randomly generated word vectors, and word vectors combined with character n-grams. The feature set was down-sized in the networks by max-pooling, and a softmax function used to classify tweets. Tested by 10-fold cross-validation, the model based on word2vec embeddings performed best, with higher precision than recall, and a 78.3% F-score.

Detecting previously unseen named entities in text is a challenging task. The paper describes how three initial classifier models were built using Conditional Random Fields (CRFs), Support Vector Machines (SVMs) and a Long Short-Term Memory (LSTM) recurrent neural network. The outputs of these three classifiers were then used as features to train another CRF classifier working as an ensemble. 5-fold cross-validation based on training and development data for the emerging and rare named entity recognition shared task showed precision, recall and F1-score of 66.87%, 46.75% and 54.97%, respectively. For surface form evaluation, the CRF ensemble-based system achieved precision, recall and F1 scores of 65.18%, 45.20% and 53.30%. When applied to unseen test data, the model reached 47.92% precision, 31.97% recall and 38.55% F1-score for entity level evaluation, with the corresponding surface form evaluation values of 44.91%, 30.47% and 36.31%.

Social media texts are often fairly informal and conversational, and when produced by bilinguals tend to be written in several different languages simultaneously, in the same way as conversational speech. The recent availability of large social media corpora has thus also made large-scale code-switched resources available for research. The paper addresses the issues of evaluation and comparison these new corpora entail, by defining an objective measure of corpus level complexity of code-switched texts. It is also shown how this formal measure can be used in practice, by applying it to several code-switched corpora.

Twitter named entity recognition is the process of identifying proper names and classifying them into some predefined labels/categories. The paper introduces a Twitter named entity system using a supervised machine learning approach, namely Conditional Random Fields. A large set of different features was developed and the system was trained using these. The Twitter named entity task can be divided into two parts: i) Named entity extraction from tweets and ii) Twitter name classification into ten different types. For Twitter named entity recognition on unseen test data, our system obtained the second highest F1 score in the shared task: 63.22%. The system performance on the classification task was worse, with an F1 measure of 40.06% on unseen test data, which was the fourth best of the ten systems participating in the shared task.