FACTOR talk: Ranasinghe – Offensive or Not? Are we doing offensive language detection in the correct way?
Thursday 6 March 2025, 3:00pm to 4:00pm
Venue
Bowland North SR 10, Lancaster, LA1 4YNOpen to
All Lancaster University (non-partner) students, Alumni, Applicants, External Organisations, Families and young people, Postgraduates, Prospective International Students, Prospective Postgraduate Students, Prospective Undergraduate Students, Public, Staff, UndergraduatesRegistration
Registration not required - just turn upEvent Details
Offensive speech online remains a challenge, partly due to differing views on what is offensive. This talk explores content moderation in political discussions, human vs. AI disagreements, and ways to improve moderation research. It also examines integrating these insights into subjective NLP tasks.
Offensive speech on web platforms is a persistent problem with wide-ranging impacts. There are many automatic content moderation systems that detect offensive content on web platforms. However, among many reasons why these tools fail to eliminate the problem is the reality that people often disagree on what is offensive.
In this talk, I will talk about how content moderation issues arise in political discussions, how machine and human moderators disagree on what is offensive when it comes to real-world social web political discourse and how content moderation research should address them. I will then present my broader vision of integrating this idea into other subjective natural language processing tasks.
Speaker
School of Computing and Communications, Lancaster University
My research focuses on developing Machine Learning (ML) approaches for Natural Language Processing (NLP) tasks, particularly emphasising NLP for Social Good. My work has various applications such as machine translation, social media analytics (i.e. hate speech detection, misinformation detection) and text adaptation. I served as an area chair in LREC-COLING 2024. I am in the programming committee of major NLP/ ML conferences. I served as the research technical track chair in EAMT 2023. I am the
Contact Details
| Name | Claire Hardaker |
| Website |