In the digital age, cyberbullying (CB) poses a significant concern, impacting individuals as early as primary school and leading to severe or lasting consequences, including an increased risk of self-harm. CB incidents, are not limited to bullies and victims, but include bystanders with various roles, and usually have numerous sub-categories and variations of online harms. This position paper emphasises the complexity of CB incidents by drawing on insights from psychology, social sciences, and computational linguistics. While awareness of CB complexities is growing, existing computational techniques tend to oversimplify CB as a binary classification task, often relying on training datasets that capture peripheries of CB behaviours. Inconsistent definitions and categories of CB-related online harms across various platforms further complicates the issue. Ethical concerns arise when CB research involves children to role-play CB incidents to curate datasets. Through multi-disciplinary collaboration, we propose strategies for consideration when developing CB detection systems. We present our position on leveraging large language models (LLMs) such as Claude-2 and Llama2-Chat as an alternative approach to generate CB-related role-playing datasets. Our goal is to assist researchers, policymakers, and online platforms in making informed decisions regarding the automation of CB incident detection and intervention. By addressing these complexities, our research contributes to a more nuanced and effective approach to combating CB especially in young people.

We conduct a comparison of pre-trained encoder-only and decoder-only language models with and without continued pre-training, to detect online sexism. Our fine-tuning-based classifier system achieved the 16th rank in the SemEval 2023 Shared Task 10 Subtask A that asks to distinguish sexist and non-sexist texts. Additionally, we conduct experiments aimed at enhancing the interpretability of systems designed to detect online sexism. Our findings provide insights into the features and decision-making processes underlying our classifier system, thereby contributing to a broader effort to develop explainable AI models to detect online sexism.

We examine the behaviour of an aspect-based sentiment classifier built by fine-tuning the BERT BASE model on the SemEval 2016 English dataset. In a set of masking experiments, we examine the extent to which the tokens identified as salient by LIME and a gradient-based method are being used by the classifier. We find that both methods are able to produce faithful rationales, with LIME outperforming the gradient-based method. We also identify a set of manually annotated sentiment expressions for this dataset, and carry out more masking experiments with these as human rationales. The enhanced performance of a classifier that only sees the relevant sentiment expressions suggests that they are not being used to their full potential. A comparison of the LIME and gradient rationales with the sentiment expressions reveals only a moderate level of agreement. Some disagreements are related to the fixed length of the rationales and the tendency of the rationales to contain content words related to the aspect itself.

The BERT family of neural language models have become highly popular due to their ability to provide sequences of text with rich context-sensitive token encodings which are able to generalise well to many NLP tasks. We introduce gaBERT, a monolingual BERT model for the Irish language. We compare our gaBERT model to multilingual BERT and the monolingual Irish WikiBERT, and we show that gaBERT provides better representations for a downstream parsing task. We also show how different filtering criteria, vocabulary size and the choice of subword tokenisation model affect downstream performance. We compare the results of fine-tuning a gaBERT model with an mBERT model for the task of identifying verbal multiword expressions, and show that the fine-tuned gaBERT model also performs better at this task. We release gaBERT and related code to the community.

We describe the DCU-EPFL submission to the IWPT 2021 Parsing Shared Task: From Raw Text to Enhanced Universal Dependencies. The task involves parsing Enhanced UD graphs, which are an extension of the basic dependency trees designed to be more facilitative towards representing semantic structure. Evaluation is carried out on 29 treebanks in 17 languages and participants are required to parse the data from each language starting from raw strings. Our approach uses the Stanza pipeline to preprocess the text files, XLM-RoBERTa to obtain contextualized token representations, and an edge-scoring and labeling model to predict the enhanced graph. Finally, we run a postprocessing script to ensure all of our outputs are valid Enhanced UD graphs. Our system places 6th out of 9 participants with a coarse Enhanced Labeled Attachment Score (ELAS) of 83.57. We carry out additional post-deadline experiments which include using Trankit for pre-processing, XLM-RoBERTa LARGE, treebank concatenation, and multitask learning between a basic and an enhanced dependency parser. All of these modifications improve our initial score and our final system has a coarse ELAS of 88.04.

We describe two Jupyter notebooks that form the basis of two assignments in an introductory Natural Language Processing (NLP) module taught to final year undergraduate students at Dublin City University. The notebooks show the students how to train a bag-of-words polarity classifier using multinomial Naive Bayes, and how to fine-tune a polarity classifier using BERT. The students take the code as a starting point for their own experiments.

We compare two orthogonal semi-supervised learning techniques, namely tri-training and pretrained word embeddings, in the task of dependency parsing. We explore language-specific FastText and ELMo embeddings and multilingual BERT embeddings. We focus on a low resource scenario as semi-supervised learning can be expected to have the most impact here. Based on treebank size and available ELMo models, we select Hungarian, Uyghur (a zero-shot language for mBERT) and Vietnamese. Furthermore, we include English in a simulated low-resource setting. We find that pretrained word embeddings make more effective use of unlabelled data than tri-training but that the two approaches can be successfully combined.

A recent advance in monolingual dependency parsing is the idea of a treebank embedding vector, which allows all treebanks for a particular language to be used as training data while at the same time allowing the model to prefer training data from one treebank over others and to select the preferred treebank at test time. We build on this idea by 1) introducing a method to predict a treebank vector for sentences that do not come from a treebank used in training, and 2) exploring what happens when we move away from predefined treebank embedding vectors during test time and instead devise tailored interpolations. We show that 1) there are interpolated vectors that are superior to the predefined ones, and 2) treebank vectors can be predicted with sufficient accuracy, for nine out of ten test languages, to match the performance of an oracle approach that knows the most suitable predefined treebank embedding for the test set.

We describe the ADAPT system for the 2020 IWPT Shared Task on parsing enhanced Universal Dependencies in 17 languages. We implement a pipeline approach using UDPipe and UDPipe-future to provide initial levels of annotation. The enhanced dependency graph is either produced by a graph-based semantic dependency parser or is built from the basic tree using a small set of heuristics. Our results show that, for the majority of languages, a semantic dependency parser can be successfully applied to the task of parsing enhanced dependencies. Unfortunately, we did not ensure a connected graph as part of our pipeline approach and our competition submission relied on a last-minute fix to pass the validation script which harmed our official evaluation scores significantly. Our submission ranked eighth in the official evaluation with a macro-averaged coarse ELAS F1 of 67.23 and a treebank average of 67.49. We later implemented our own graph-connecting fix which resulted in a score of 79.53 (language average) or 79.76 (treebank average), which would have placed fourth in the competition evaluation.

Cross-lingual dependency parsing involves transferring syntactic knowledge from one language to another. It is a crucial component for inducing dependency parsers in low-resource scenarios where no training data for a language exists. Using Faroese as the target language, we compare two approaches using annotation projection: first, projecting from multiple monolingual source models; second, projecting from a single polyglot model which is trained on the combination of all source languages. Furthermore, we reproduce multi-source projection (Tyers et al., 2018), in which dependency trees of multiple sources are combined. Finally, we apply multi-treebank modelling to the projected treebanks, in addition to or alternatively to polyglot modelling on the source side. We find that polyglot training on the source languages produces an overall trend of better results on the target language but the single best result for the target language is obtained by projecting from monolingual source parsing models and then training multi-treebank POS tagging and parsing models on the target side.

Automatic post-editing (APE) can be reduced to a machine translation (MT) task, where the source is the output of a specific MT system and the target is its post-edited variant. However, this approach does not consider context information that can be found in the original source of the MT system. Thus a better approach is to employ multi-source MT, where two input sequences are considered – the one being the original source and the other being the MT output. Extra context information can be introduced in the form of extra tokens that identify certain global property of a group of segments, added as a prefix or a suffix to each segment. Successfully applied in domain adaptation of MT as well as on APE, this technique deserves further attention. In this work we investigate multi-source neural APE (or NPE) systems with training data which has been augmented with two types of extra context tokens. We experiment with authentic and synthetic data provided by WMT 2019 and submit our results to the APE shared task. We also experiment with using statistical machine translation (SMT) methods for APE. While our systems score bellow the baseline, we consider this work a step towards understanding the added value of extra context in the case of APE.