Abstract

Sexism, an injustice that subjects women and girls to enormous suffering, manifests in blatant as well as subtle ways. In the wake of growing documentation of experiences of sexism on the web, the automatic categorization of accounts of sexism has the potential to assist social scientists and policy makers in utilizing such data to study and counter sexism better. The existing work on sexism classification, which is different from sexism detection, has certain limitations in terms of the categories of sexism used and/or whether they can co-occur. To the best of our knowledge, this is the first work on the multi-label classification of sexism of any kind(s), and we contribute the largest dataset for sexism categorization. We develop a neural solution for this multilabel classification that can combine sentence representations obtained using models such as BERT with distributional and linguistic word embeddings using a flexible, hierarchical architecture involving recurrent components and optional convolutional ones. Further, we leverage unlabeled accounts of sexism to infuse domain-specific elements into our framework. The best proposed method outperforms several deep learning as well as traditional machine learning baselines by an appreciable margin.

Abstract

This paper describes our system (Fermi) for Task 6: OffensEval: Identifying and Categorizing Offensive Language in Social Media of SemEval-2019. We participated in all the three sub-tasks within Task 6. We evaluate multiple sentence embeddings in conjunction with various supervised machine learning algorithms and evaluate the performance of simple yet effective embedding-ML combination algorithms. Our team Fermi’s model achieved an F1-score of 64.40%, 62.00% and 62.60% for sub-task A, B and C respectively on the official leaderboard. Our model for sub-task C which uses pre-trained ELMo embeddings for transforming the input and uses SVM (RBF kernel) for training, scored third position on the official leaderboard. Through the paper we provide a detailed description of the approach, as well as the results obtained for the task.

Abstract

This paper describes our system (Fermi) for Task 5 of SemEval-2019: HatEval: Multilingual Detection of Hate Speech Against Immigrants and Women on Twitter. We participated in the subtask A for English and ranked first in the evaluation on the test set. We evaluate the quality of multiple sentence embeddings and explore multiple training models to evaluate the performance of simple yet effective embedding-ML combination algorithms. Our team-Fermi’s model achieved an accuracy of 65.00% for English language in task A. Our models, which use pretrained Universal Encoder sentence embeddings for transforming the input and SVM (with RBF kernel) for classification, scored first position (among 68) in the leaderboard on the test set for Subtask A in English language. In this paper we provide a detailed description of the approach, as well as the results obtained in the task.

Abstract

Online Community Question Answering Forums (cQA) have gained massive popularity within recent years. The rise in users for such forums have led to the increase in the need for automated evaluation for question comprehension and fact evaluation of the answers provided by various participants in the forum. Our team, Fermi, participated in sub-task A of Task 8 at SemEval 2019-which tackles the first problem in the pipeline of factual evaluation in cQA forums, ie, deciding whether a posed question asks for a factual information, an opinion/advice or is just socializing. This information is highly useful in segregating factual questions from non-factual ones which highly helps in organizing the questions into useful categories and trims down the problem space for the next task in the pipeline for fact evaluation among the available answers. Our system uses the embeddings obtained from Universal Sentence Encoder combined with XGBoost for the classification sub-task A. We also evaluate other combinations of embeddings and off-the-shelf machine learning algorithms to demonstrate the efficacy of the various representations and their combinations. Our results across the evaluation test set gave an accuracy of 84% and received the first position in the final standings judged by the organizers.

Abstract

Modeling affect via understanding the social constructs behind them is an important task in devising robust and accurate systems for socially relevant scenarios. In the CL-Aff Shared Task (part of Affective Content Analysis workshop @ AAAI 2019), the organizers released a dataset of ‘happy’ moments, called the HappyDB corpus. The task is to detect two social constructs: the agency (i.e., whether the author is in control of the happy moment) and the social characteristics (i.e., whether anyone else other than the author was also involved in the happy moment). We employ an inductive transfer learning technique where we utilize a pre-trained language model and fine-tune it on the target task for both the binary classification tasks. At first, we use a language model pre-trained on the huge WikiText-103 corpus. This step utilizes an AWDLSTM with three hidden layers for training the language model. In the second step, we fine-tune the pre-trained language model on both the labeled and unlabeled instances from the HappyDB dataset. Finally, we train a classifier on top of the language model for each of the identification tasks. Our experiments using 10-fold cross validation on the corpus show that we achieve a high accuracy of ∼93% for detection of the social characteristic and ∼87% for agency of the author, showing significant gains over other baselines. We also show that using the unlabeled dataset for fine-tuning the language model in the second step improves our accuracy by 1-2% across detection of both the constructs.

Abstract

Cross-lingual summarization (CLS) aims to create summaries in a target language, from a document or document set given in a different, source language. Cross-lingual summarization can play a critical role in enabling cross-lingual information access for millions of people across the globe who do not speak or understand languages having large representation on the web. It can also make documents originally published in local languages quickly accessible to a large audience which does not understand those local languages. Though cross-lingual summarization has gathered some attention in the last decade, there has been no serious effort to publish rigorous software for this task. In this paper, we provide a design for an end-to-end CLS software called clstk. Besides implementing a number of methods proposed by different CLS researchers over years, the software integrates multiple components critical for CLS. We hope that this extremely modular tool-kit will help CLS researchers to contribute more effectively to the area.

Abstract

Users have been trained to type keyword queries on search engines. However, recently there has been a significant rise in the number of verbose queries. Often times such queries are not well-formed. The lack of well-formedness in the query might adversely impact the downstream pipeline which processes these queries. A well-formed natural language question as a search query aids heavily in reducing errors in downstream tasks and further helps in improved query understanding. In this paper, we employ an inductive transfer learning technique by fine-tuning a pretrained language model to identify whether a search query is a well-formed natural language question or not. We show that our model trained on a recently released benchmark dataset spanning 25,100 queries gives an accuracy of 75.03% thereby improving by ∼5 absolute percentage points over the state-of-the-art.

Abstract

Sentence compression has traditionally been tackled as syntactic tree pruning, where rules and statistical features are defined for pruning less relevant words. Recent years have witnessed the rise of neural models without leveraging syntax trees, learning sentence representations automatically and pruning words from such representations. We investigate syntax tree based noise pruning methods for neural sentence compression. Our method identifies the most informative regions in a syntactic dependency tree by self attention over context nodes and maximum density subtree extraction. Empirical results show that the model outperforms the state-of-the-art methods in terms of both accuracy and F1-measure. The model also yields a comparable accuracy in readability and informativeness as assessed by human evaluators.

Abstract

Spatial Role Labelling involves identification of text segments which emit spatial semantics such as describing an object of interest, a reference point or the object’s relative position with the reference. Tasks in SemEval exercises of 2012 and 2013 propose problems and datasets for Spatial Role Labelling. In this paper, we propose a simple two-step neural network based approach to identify static spatial relations along with the three primary roles - Trajector, Landmark and Spatial Indicator. Our approach outperforms the task submission results and other state-of-the-art results on these datasets. We also include a discussion on the explainability of our model.

Abstract

With the ever-increasing cases of hate spread on social media platforms, it is critical to design abuse detection mechanisms to proactively avoid and control such incidents. While there exist methods for hate speech detection, they stereotype words and hence suffer from inherently biased training. Bias removal has been traditionally studied for structured datasets, but we aim at bias mitigation from unstructured text data. In this paper, we make two important contributions. First, we systematically design methods to quantify the bias for any model and propose algorithms for identifying the set of words which the model stereotypes. Second, we propose novel methods leveraging knowledge-based generalizations for bias-free learning. Knowledge-based generalization provides an effective way to encode knowledge because the abstraction they provide not only generalizes content but also facilitates retraction of information from the hate speech detection classifier, thereby reducing the imbalance. We experiment with multiple knowledge generalization policies and analyze their effect on general performance and in mitigating bias. Our experiments with two real-world datasets, a Wikipedia Talk Pages dataset (WikiDetox) of size ∼96k and a Twitter dataset of size ∼24k, show that the use of knowledge-based generalizations results in better performance by forcing the classifier to learn from generalized content. Our methods utilize existing knowledge-bases and can easily be extended to other tasks.