In today's digital era, social media have become primary platforms for disseminating newsworthy content, with most internet users relying on them for regular updates. Thus, understanding and detecting important events from social media data streams is vital for various applications ranging from crisis management to market analysis. However, the vast volume and unstructured nature of this data, generated by a diverse range of users, make manual detection methods highly labour-intensive. As a result, automated intelligent mechanisms have become essential for efficiently handling event detection tasks. However, most available social media event detection approaches primarily rely on data statistics, ignoring semantics, making them vulnerable to critical information loss. Following this gap, this presentation will explore how temporal word dynamics can be involved in achieving effective event detection from social media data.

Thursday, December 5, 2024

15:00 - 16:00

Bowland North SR02

Dr Hansi Hettiarachchi

Talk / Public Lecture

Events - Campus and Community Life, Events - Doctoral Academy, Events - Staff Channel, Events - Student Channel, Library Channel, Web - Embrace Digital Staff, Web - Embrace Digital Students

This talk will be given by Dr Hansi Hettiarachchi (Computing, SPS, Lancaster University). Hansi's research primarily focuses on developing Machine Learning (ML) approaches for Natural Language Processing (NLP) tasks, particularly emphasising societal and human security and safety. Her recent work has concentrated on three main areas: online event detection (including temporal and textual event profiling, trigger and argument detection, and event causality identification), online safety (covering offensive content detection, misinformation identification, and fake news detection), and information extraction (such as named entity recognition, relation extraction, and rule generation). The recent advancements in NLP have significantly impacted society and human beings, providing many benefits while raising serious concerns about the trustworthiness, truthfulness, and fairness of NLP technologies. Therefore, she is also interested in exploring the capabilities of Large Language Models (LLMs), the multilingualism of models/NLP technologies, support for low-resource languages, and the explainability of models/ NLP technologies.