Abstract

Memes, a prevalent form of humour in the digital era, efficiently convey sarcasm, wit, and cultural references, spreading rapidly on the internet due to their relatability and shareable nature, exacerbating conflicts. While memes offer insights into online communities and the dynamics of information diffusion, they also pose risks by spreading disinformation and misinformation, hate speech, cyberbullying, and harmful content.  In this talk, I will discuss our work on the Memology project, which aims to enhance our understanding and decipher the language of internet memes. The work focuses on developing advanced computational methods for large-scale meme analysis.  Specifically, I will present three works that showcase the creation of novel algorithms to identify misinformation and harmful content within memes. Concluding the talk, I will share the results that demonstrate the efficacy and adaptability of these models and propose future directions for further research.

About the Speaker

Usman Naseem is a Lecturer (equivalent to Assistant Professor) in the School of Computing at Macquarie University, Australia. Before transitioning to academia, he worked in industry for over 10 years in various technical and leadership roles. His research interests include natural language processing, multimodal analysis, and social computing. Specifically, he focuses on designing socially aware innovative methods for various applications.