Social media is becoming a primary medium to discuss what is happening around the world. Therefore, the data generated by social media platforms contain rich information which describes the ongoing events. Further, the timeliness associated with these data is capable of facilitating immediate insights. However, considering the dynamic nature and high volume of data production in social media data streams, it is impractical to filter the events manually and therefore, automated event detection mechanisms are invaluable to the community. Apart from a few notable exceptions, most previous research on automated event detection have focused only on statistical and syntactical features in data and lacked the involvement of underlying semantics which are important for effective information retrieval from text since they represent the connections between words and their meanings. In this paper, we propose a novel method termedEmbed2Detectfor event detection in social media by combining the characteristics in word embeddings and hierarchical agglomerative clustering. The adoption of word embeddings givesEmbed2Detectthe capability to incorporate powerful semantical features into event detection and overcome a major limitation inherent in previous approaches. We experimented our method on two recent real social media data sets which represent the sports and political domain and also compared the results to several state-of-the-art methods. The obtained results show thatEmbed2Detectis capable of effective and efficient event detection and it outperforms the recent event detection methods. For the sports data set, Embed2Detect achieved 27% higher F-measure than the best-performed baseline and for the political data set, it was an increase of 29%.

Automatic Compliance Checking (ACC) within the Architecture, Engineering, and Construction (AEC) sector necessitates automating the interpretation of building regulations to achieve its full potential. Converting textual rules into machine-readable formats is challenging due to the complexities of natural language and the scarcity of resources for advanced Machine Learning (ML). Addressing these challenges, we introduce CODE-ACCORD, a dataset of 862 sentences from the building regulations of England and Finland. Only the self-contained sentences, which express complete rules without needing additional context, were considered as they are essential for ACC. Each sentence was manually annotated with entities and relations by a team of 12 annotators to facilitate machine-readable rule generation, followed by careful curation to ensure accuracy. The final dataset comprises 4,297 entities and 4,329 relations across various categories, serving as a robust ground truth. CODE-ACCORD supports a range of ML and Natural Language Processing (NLP) tasks, including text classification, entity recognition, and relation extraction. It enables applying recent trends, such as deep neural networks and large language models, to ACC.

The COVID-19 pandemic has brought economic activity to a near standstill as many countries imposed very strict restrictions on movement to halt the spread of the virus. This study aims at assessing the economic impacts of COVID-19 in the United Kingdom (UK) using artificial intelligence (AI) and data from previous economic crises to predict future economic impacts. The macroeconomic indicators, gross domestic products (GDP) and GDP growth, and data on the performance of three primary industries in the UK (the construction, production and service industries) were analysed using a comparison with the pattern of previous economic crises. In this research, we experimented with the effectiveness of both continuous and categorical time-series forecasting on predicting future values to generate more accurate and useful results in the economic domain. Continuous value predictions indicate that GDP growth in 2021 will remain steady, but at around −8.5% contraction, compared to the baseline figures before the pandemic. Further, the categorical predictions indicate that there will be no quarterly drop in GDP following the first quarter of 2021. This study provided evidence-based data on the economic effects of COVID-19 that can be used to plan necessary recovery procedures and to take appropriate actions to support the economy.

Today, we have access to a vast data amount, especially on the internet. Online news agencies play a vital role in this data generation, but most of their data is unstructured, requiring an enormous effort to extract important information. Thus, automated intelligent event detection mechanisms are invaluable to the community. In this research, we focus on identifying event details at the sentence and token levels from news articles, considering their fine granularity. Previous research has proposed various approaches ranging from traditional machine learning to deep learning, targeting event detection at these levels. Among these approaches, transformer-based approaches performed best, utilising transformers’ transferability and context awareness, and achieved state-of-the-art results. However, they considered sentence and token level tasks as separate tasks even though their interconnections can be utilised for mutual task improvements. To fill this gap, we propose a novel learning strategy named Two-phase Transfer Learning (TTL) based on transformers, which allows the model to utilise the knowledge from a task at a particular data granularity for another task at different data granularity, and evaluate its performance in sentence and token level event detection. Also, we empirically evaluate how the event detection performance can be improved for different languages (high- and low-resource), involving monolingual and multilingual pre-trained transformers and language-based learning strategies along with the proposed learning strategy. Our findings mainly indicate the effectiveness of multilingual models in low-resource language event detection. Also, TTL can further improve model performance, depending on the involved tasks’ learning order and their relatedness concerning final predictions.

Even though social media contain rich information on events and public opinions, it is impractical to manually filter this information due to data’s vast generation and dynamicity. Thus, automated extraction mechanisms are invaluable to the community. We need real data with ground truth labels to build/evaluate such systems. Still, to the best of our knowledge, no available social media dataset covers continuous periods with event and sentiment labels together except for events or sentiments. Datasets without time gaps are huge due to high data generation and require extensive effort for manual labelling. Different approaches, ranging from unsupervised to supervised, have been proposed by previous research targeting such datasets. However, their generic nature mainly fails to capture event-specific sentiment expressions, making them inappropriate for labelling event sentiments. Filling this gap, we propose a novel data annotation approach in this paper involving several neural networks. Our approach outperforms the commonly used sentiment annotation models such as VADER and TextBlob. Also, it generates probability values for all sentiment categories besides providing a single category per tweet, supporting aggregated sentiment analyses. Using this approach, we annotate and release a dataset named TED-S, covering two diverse domains, sports and politics. TED-S has complete subsets of Twitter data streams with both sub-event and sentiment labels, providing the ability to support event sentiment-based research.

Tharindu Ranasinghe, Hansi Hettiarachchi, Constantin Orasan, Ruslan Mitkov. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). 2025.