News
Winners of our inaugural CIRES Best Demo Award!
We are pleased to announce the winners of our inaugural CIRES Best Demo Award! Congratulations to PhD Researchers Elyas Meguellati and Stefano Civelli from The University of Queensland who received a $1,000 AUD prize for their demo, the Ad Persuasion Dashboard – Insights into Facebook Political Campaign Strategies. The demo presents an interactive dashboard to […]
We are pleased to announce the winners of our inaugural CIRES Best Demo Award!
Congratulations to PhD Researchers Elyas Meguellati and Stefano Civelli from The University of Queensland who received a $1,000 AUD prize for their demo, the Ad Persuasion Dashboard – Insights into Facebook Political Campaign Strategies.
The demo presents an interactive dashboard to analyse persuasive content in political advertising on social media. Focusing on Facebook ads from the 2022 Australian Federal Election campaign, it uses a state-of-the-art lightweight model for persuasive text detection. The web application allows users to gain insights through visualisations of comparative spend and impressions on high vs. low persuasion ads, time series analysis of ad impressions, and demographic targeting patterns. The tool enhances transparency in digital pollical campaigning by enabling researchers and the public to explore persuasion strategies employed in social media advertising.
This work has been submitted to the ACM Web Conference (WWW) 2025. Future applications and work identified include more analyses to be shown and the ability for other researchers to upload their dataset.
Elyas’s PhD project, titled ‘The Duality of Persuasion,’ delves into both the generation and detection aspects of persuasive communication. On the generation side, his research focuses on creating tailored messages that align with specific personality traits, while the detection side emphasizes identifying persuasive techniques in textual content. He is supervised by Profs. Gianluca Demartini and Shazia Sadiq.
Stefano’s PhD research is focused on developing novel methodologies for measuring and understanding prompt complexity in Large Language Models (LLMs). His work aims to identify and quantify the key factors that contribute to prompt complexity, with practical applications ranging from optimal model selection to response quality prediction. Working under the supervision of Prof. Gianluca Demartini, his research aims to advance our understanding of how to more effectively interact with and deploy LLMs in real-world applications
Find out more:
Pictured L to R: Elyas Meguellati, Prof. Shazia Sadiq, Dr Junliang Yu, & Stefano Civelli at CIRES HQ.
