In this paper, we studied the gender bias in monolingual word embeddings of two Indian languages Hindi and Tamil. Tamil is one of the classical languages of India from the Dravidian language family. In Indian society and culture, instead of racism, a similar type of discrimination called casteism is against the subgroup of peoples representing lower class or Dalits. The word embeddings measurement to evaluate bias using the WEAT score reveals that the embeddings are biased with gender and casteism which is in line with the common stereotypical human biases.

Gender biases in syntax have been documented for languages with grammatical gender for cases where mixed-gender coordination structures take masculine agreement, or with male-first preference in the ordering of pairs (Adam and Eve). On the basis of various annotated corpora spanning different genres (fiction, newspapers, speech and web), we show another syntactic gender bias: masculine pronouns are more often subjects than feminine pronouns, in both English and French. We find the same bias towards masculine subjects for French human nouns, which then refer to males and females. Comparing the subject of passive verbs and the object of active verbs, we show that this syntactic function bias is not reducible to a bias in semantic role assignment since it is also found with non-agentive subjects. For French fiction, we also found that the masculine syntactic function bias is larger in text written by male authors – female authors seem to be unbiased. We finally discuss two principles as possible explanations, ‘Like Me’ and ‘Easy first’, and examine the effect of the discourse tendency for men being agents and topics. We conclude by addressing the impact of such biases in language technologies.

Cancel Culture as an Internet phenomenon has been previously explored from a social and legal science perspective. This paper demonstrates how Natural Language Processing tasks can be derived from this previous work, underlying techniques on how cancel culture can be measured, identified and evaluated. As part of this paper, we introduce a first cancel culture data set with of over 2.3 million tweets and a framework to enlarge it further. We provide a detailed analysis of this data set and propose a set of features, based on various models including sentiment analysis and emotion detection that can help characterizing cancel culture.

Cyberbullying is bullying perpetrated via the medium of modern communication technologies like social media networks and gaming platforms. Unfortunately, most existing datasets focusing on cyberbullying detection or classification are i) limited in number ii) usually targeted to one specific online social networking (OSN) platform, or iii) often contain low-quality annotations. In this study, we fine-tune and benchmark state of the art neural transformers for the binary classification of cyberbullying in social media texts, which is of high value to Natural Language Processing (NLP) researchers and computational social scientists. Furthermore, this work represents the first step toward building neural language models for cross OSN platform cyberbullying classification to make them as OSN platform agnostic as possible.

Discriminatory language, in particular hate speech, is a global problem posing a grave threat to democracy and human rights. Yet, it is not always easy to identify, as it is rarely explicit. In order to detect hate speech, we developed Hierarchical Attention Network (HAN) based and Bidirectional Encoder Representations from Transformer (BERT) based deep learning models to capture the changing discursive cues and understand the context around the discourse. In addition, we designed linguistic features using critical discourse analysis techniques and integrated them into these neural network models. We studied the compatibility of our model with the hate speech detection problem by comparing it with traditional machine learning models, as well as a Convolution Neural Network (CNN) based model, a Convolutional Neural Network-Gated Recurrent Unit (CNN-GRU) based model which reached significant performance results for hate speech detection. Our results on a manually annotated corpus of print media in Turkish show that the proposed approach is effective for hate speech detection. We believe that the feature sets created for the Turkish language will encourage new studies in the quantitative analysis of hate speech.

Legal field is characterized by its exclusivity and non-transparency. Despite the frequency and relevance of legal dealings, legal documents like contracts remains elusive to non-legal professionals for the copious usage of legal jargon. There has been little advancement in making legal contracts more comprehensible. This paper presents how Machine Learning and NLP can be applied to solve this problem, further considering the challenges of applying ML to the high length of contract documents and training in a low resource environment. The largest open-source contract dataset so far, the Contract Understanding Atticus Dataset (CUAD) is utilized. Various pre-processing experiments and hyperparameter tuning have been carried out and we successfully managed to eclipse SOTA results presented for models in the CUAD dataset trained on RoBERTa-base. Our model, A-type-RoBERTa-base achieved an AUPR score of 46.6% compared to 42.6% on the original RoBERT-base. This model is utilized in our end to end contract understanding application which is able to take a contract and highlight the clauses a user is looking to find along with it’s descriptions to aid due diligence before signing. Alongside digital, i.e. searchable, contracts the system is capable of processing scanned, i.e. non-searchable, contracts using tesseract OCR. This application is aimed to not only make contract review a comprehensible process to non-legal professionals, but also to help lawyers and attorneys more efficiently review contracts.