The manifestation of cyberbullying (CB) on social media platforms (SMP) evolves alongside advancements in Internet communication technologies (ICT).Current automated CB detection techniques, often relying on artificial intelligence (AI) are overly simplistic, overlooking the multimodal nature of user-generated content and the relational nature of CB and treating it as a binary problem.Addressing these limitations, this mixed methods study involves young people (aged 13-17) in developing CB detection software following the principles of the Software Development Life Cycle (SDLC).To explore the perspectives of young people on online harms related to CB across various SMPs, we first surveyed 104 participants aged 13-17 in Ireland.Following the call of the European Union Joint Research Center (EU JRC) to integrate young people's perspectives into the design of AI systems that serve their needs, we facilitate co-design sessions with "junior researchers" (aged 15-16) during a week-long research internship.In these sessions, participants adapt existing CB scenarios to reflect the realities of today's SMP and ICT environments, ensuring the outcomes are relevant and youth-informed.Additionally, the junior researchers help craft more than 550 text messages 1 across multiple cyberbullying scenarios that reflect bystander-enabler and defender behaviours (e.g., escalate or de-escalate the cyberbullying behaviour).Tailored to reflect the dynamics of the comment-thread and mulit-party conversations, these messages are current under review by experts.This structured and generative input positions them from users to co-designers of a youth-informed benchmark dataset that will support the evaluation of future CB detection systems.

The ability of LLMs to write coherent, faithful long texts from structured data inputs remains relatively uncharted, in part because nearly all public data-to-text datasets contain only short input-output pairs. To address these gaps, we benchmark six LLMs, a rule‐based system and human-written texts on a new long-input dataset in English and Irish via LLM-based evaluation. We find substantial differences between models and languages.

Large Language Models (LLMs) have achieved remarkable results in natural language generation, yet challenges remain in data-to-text (D2T) tasks, particularly in controlling output, ensuring transparency, and maintaining factual consistency with the input. We introduce the first LLM-based multi-agent framework for D2T generation, coordinating specialized agents to produce high-quality, interpretable outputs. Our system combines the reasoning and acting abilities of ReAct agents, the self-correction of Reflexion agents, and the quality assurance of Guardrail agents, all directed by an Orchestrator agent that assigns tasks to three specialists—content ordering, text structuring, and surface realization—and iteratively refines outputs based on Guardrail feedback. This closed-loop design enables precise control and dynamic optimization, yielding text that is coherent, accurate, and grounded in the input data. On a relatively simple dataset like WebNLG, our framework performs competitively with end-to-end systems, highlighting its promise for more complex D2T scenarios.

Minority languages such as Irish are massively under-resourced, particularly in terms of high-quality domain-relevant data, limiting the capabilities of machine translation (MT) engines, even those integrating large language models (LLMs). The eSTÓR project, described in this paper, focuses on the collection and curation of high-quality Irish text data for diverse domains.

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.

Previous studies have highlighted the advantages of pipeline neural architectures over endto-end models, particularly in reducing text hallucination.In this study, we extend prior research by integrating pretrained language models (PLMs) into a pipeline framework, using both fine-tuning and prompting methods.Our findings show that fine-tuned PLMs consistently generate high quality text, especially within end-to-end architectures and at intermediate stages of the pipeline across various domains.These models also outperform promptbased ones on automatic evaluation metrics but lag in human evaluations.Compared to the standard five-stage pipeline architecture, a streamlined three-stage pipeline, which only include ordering, structuring, and surface realization, achieves superior performance in fluency and semantic adequacy according to the human evaluation.

Abstract Cyberbullying is the wilful and repeated infliction of harm on an individual using the Internet and digital technologies. Similar to face-to-face bullying, cyberbullying can be captured formally using the Routine Activities Model (RAM) whereby the potential victim and bully are brought into proximity of one another via the interaction on online social networking (OSN) platforms. Although the impact of the COVID-19 (SARS-CoV-2) restrictions on the online presence of minors has yet to be fully grasped, studies have reported that 44% of pre-adolescents have encountered more cyberbullying incidents during the COVID-19 lockdown. Transparency reports shared by OSN companies indicate an increased take-downs of cyberbullying-related comments, posts or content by artificially intelligen moderation tools. However, in order to efficiently and effectively detect or identify whether a social media post or comment qualifies as cyberbullying, there are a number factors based on the RAM, which must be taken into account, which includes the identification of cyberbullying roles and forms. This demands the acquisition of large amounts of fine-grained annotated data which is costly and ethically challenging to produce. In addition where fine-grained datasets do exist they may be unavailable in the target language. Manual translation is costly and expensive, however, state-of-the-art neural machine translation offers a workaround. This study presents a first of its kind experiment in leveraging machine translation to automatically translate a unique pre-adolescent cyberbullying gold standard dataset in Italian with fine-grained annotations into English for training and testing a native binary classifier for pre-adolescent cyberbullying. In addition to contributing high-quality English reference translation of the source gold standard, our experiments indicate that the performance of our target binary classifier when trained on machine-translated English output is on par with the source (Italian) classifier.

This paper reports results from a reproduction study in which we repeated the human evaluation of the PASS Dutch-language football report generation system (van der Lee et al., 2017).The work was carried out as part of the ReproGen Shared Task on Reproducibility of Human Evaluations in NLG, in Track A (Paper 1).We aimed to repeat the original study exactly, with the main difference that a different set of evaluators was used.We describe the study design, present the results from the original and the reproduction study, and then compare and analyse the differences between the two sets of results.For the two 'headline' results of average Fluency and Clarity, we find that in both studies, the system was rated more highly for Clarity than for Fluency, and Clarity had higher standard deviation.Clarity and Fluency ratings were higher, and their standard deviations lower, in the reproduction study than in the original study by substantial margins.Clarity had a higher degree of reproducibility than Fluency, as measured by the coefficient of variation.Data and code are publicly available.1

Knowledge base creation and population are an essential formal backbone for a variety of intelligent applications, decision support and expert systems and intelligent search. While the abundance of unstructured text helps in easing the knowledge acquisition gap, the ambiguous nature of language tends to impact accuracy when engaging in more complex semantic analysis. Controlled Natural Languages (CNLs) are subsets of natural language that are restricted grammatically in order to reduce or eliminate ambiguity for the purposes of machine processability, or unambiguous human communication within a domain or industry context, such as Simplified English. This type of human-oriented CNL is under-researched despite having found favor within industry over many years. We describe a novel dataset which aligns a representative sample of Simplified English Wikipedia sentences with a well known machine-oriented CNL. This linguistic resource is both human-readable and semantically machine interpretable and can benefit a variety of NLP and knowledge based applications.

Different semantic interpretation tasks such as text entailment and question answering require the classification of semantic relations between terms or entities within text. However, in most cases, it is not possible to assign a direct semantic relation between entities/terms. This paper proposes an approach for composite semantic relation classification using one or more relations between entities/term mentions, extending the traditional semantic relation classification task . The proposed model is different from existing approaches which typically use machine learning models built over lexical and distributional word vector features in that is uses a combination of a large commonsense knowledge base of binary relations , a distributional navigational algorithm and sequence classification to provide a solution for the composite semantic relation classification problem. The proposed approach outperformed existing baselines with regard to F1-score, Accuracy, Precision and Recall.

The growth in the alternative digital publishing is widening the breadth of scholarly impact beyond the conventional bibliometric community. Thus, research is becoming more reachable both inside and outside of academic institutions and are found to be shared, downloaded and discussed in social media. In this study, we linked the scientific articles found in mainstream news, weblogs and Stack Overflow to the citation database of peer-reviewed literature called Scopus. We then explored how standard graph-based influence metrics can be used to measure the social impact of scientific articles. We also proposed the variant of Katz centrality metrics called EgoMet score to measure the local importance of scientific articles in its ego network. Later we evaluated these computed graph-based influence metrics by predicting absolute citations. Our results of the prediction model describe 34% variance to predict citations from blogs and mainstream news and 44% variance to predict citations from Stack Overflow.

The realization that scholarly publications are discussed and have influence on discourse outside scientific and academic domains has given rise to area of scientometrics called alternative metrics or “altmetrics”. Furthermore, researchers in this field tend to focus primarily on measuring scientific activity on social media platforms such as Twitter, however these count-based metrics are vulnerable to gaming because they tend to lack concrete justification or reference to the primary source. In this collaboration with Elsevier, we extend the conventional citation graph to a heterogeneous graph of publications, scientists, venues, organizations and more authoritative media sources such as mainstream news and weblogs. Our approach consists of two parts: one is integrating the bibliometric data with the social data such as blogs, mainstream news. The other involves understanding how standard graph-based metrics can be used to predict the academic impact. Our result showed the computed graph-based metrics can reasonably predict the academic impact of early stage researchers.

The goal of this paper is to propose a system that can extract formal semantic knowledge representation from natural language eGov policies. We present an architecture that allows for extracting Controlled Natural Language (CNL) statements from heterogeneous natural language texts with the ability to support multilinguality. The approach is based on the concept of embedded CNLs.

The logging of Activities of Daily Living (ADLs) is becoming increasingly popular mainly thanks to wearable devices. Currently, most sensors used for ADLs logging are queried and filtered mainly by location and time. However, in an Internet of Things future, a query will return a large amount of sensor data. Therefore, existing approaches will not be feasible because of resource constraints and performance issues. Hence more fine-grained queries will be necessary. We propose to filter on the likelihood that a sensor is relevant for the currently sensed activity. Our aim is to improve system efficiency by reducing the amount of data to query, store and process by identifying which sensors are relevant for different activities during the ADLs logging by relying on Distributional Semantics over public text corpora and unsupervised hierarchical clustering. We have evaluated our system over a public dataset for activity recognition and compared our clusters of sensors with the sensors involved in the logging of manually-annotated activities. Our results show an average precision of 89% and an overall accuracy of 69%, thus outperforming the state of the art by 5% and 32% respectively. To support the uptake of our approach and to allow replication of our experiments, a Web service has been developed and open sourced.

This demonstration presents an infrastructure for computing multilingual semantic relatedness and correlation for twelve natural languages by using three distributional semantic models (DSMs). Our demonsrator - DInfra (Distributional Infrastructure) provides researchers and developers with a highly useful platform for processing large-scale corpora and conducting experiments with distributional semantics. We integrate several multilingual DSMs in our webservice so end user can obtain a result without worrying about the complexities involved in building DSMs. Our webservice allows the users to have easy access to a wide range of comparisons of DSMs with different parameters. In addition, users can configure and access DSM parameters using a easy to use API.

Entering 'Football Players from United States' when searching for 'American Footballers' is an example of vocabulary mismatch, which occurs when different words are used to express the same concepts. In order to address this phenomenon for entity search targeting descriptors for complex categories, we propose a compositional-distributional semantics entity search engine, which extracts semantic and commonsense knowledge from large-scale corpora to address the vocabulary gap between query and data.

Richly interlinked metadata constitute the foundation of the Semantic Web. Manual semantic annotation is a labor intensive task requiring training in formal ontological descriptions for the otherwise non-expert user. Although automatic annotation tools attempt to ease this knowledge acquisition barrier, their development often requires access to specialists in Natural Language Processing (NLP). This challenges researchers to develop user-friendly annotation environments. Controlled Natural Languages (CNLs) offer an incentive to the novice user to annotate, while simultaneously authoring his/her respective documents in a user-friendly manner. CNLs have been successfully applied to ontology authoring, but little research has focused on their application to semantic annotation. This paper describes two novel approaches to semantic annotation, which permit non-expert users to simultaneously author and annotate meeting minutes using CNL. Finally, this work provides empirical evidence that for certain scenarios applying CNLs for semantic annotation can be more user friendly than a standard manual semantic annotation tool.