Agostina Calabrese
(she/her)
About
I am a fourth-year Ph.D. student at the UKRI Centre for Doctoral Training in NLP (The University of Edinburgh), where I am working on policy-aware explainable abuse detection under the supervision of Björn Ross and Mirella Lapata. My research focuses on developing models that align with platform-specific moderation policies while remaining transparent and explainable, and it demonstrates how this approach can significantly reduce costs in real-world moderation pipelines. I am a member of the EdinburghNLP group based at the Institute for Language, Cognition and Computation (ILCC).
I have worked as a Research Intern in the Computational Social Science team at Snap Inc. (Fall 2023) and in the Assistant AI team at Meta Reality Labs (Fall 2022).
Previously I was a Research Fellow at Sapienza NLP (Sapienza University of Rome), where I worked on multimodal systems for language-vision tasks, and at the development of BabelPic. I hold a Bachelor’s (2017) and a Master’s degree (2019) in Computer Science from Sapienza University of Rome, both awarded with full marks and honors (110/110 cum laude).
I am also actively involved in the NLP community around online safety. I co-organised the 8th and 9th Workshop on Online Abuse and Harms (2024, 2025), and in 2022, I created a Slack community for researchers and practitioners working on hate speech and related topics. As of June 2025, it includes over 250 members from academia and industry.
I am currently on the job market, looking for Research Scientist roles in NLP, ideally with a focus on safety.
News
2025:
- My latest paper, “Compositional Generalisation for Explainable Hate Speech Detection”, is now available as a preprint on arXiv.
2024:
- I’m co-organising the 9th Workshop on Online Abuse and Harms! WOAH 9 will be at ACL 2025, and the special theme for this year is “Harms Beyond Hate Speech”.
- Our paper “Explainability and Hate Speech: Structured Explanations Make Social Media Moderators Faster” has been accepted at ACL 2024!
- I’m co-organising the 8th Workshop on Online Abuse and Harms (WOAH) - co-located with NAACL 2024.
2023:
- We are organising the 2nd Workshop on Novel Evaluation Approaches for Text Classification Systems - co-located with ICWSM 2023!
- Aug-Dec: I’m in LA working as a research intern at Snap Inc.!
2022:
- Our paper “Explainable Abuse Detection as Intent Classification and Slot Filling” has been accepted at TACL!
- Aug-Dec: I’m in the Seattle area working as a research intern at Meta AI!
- I am co-organising the 1st Workshop on Novel Evaluation Approaches for Text Classification Systems on Social Media: check out the call for papers!
2021:
- The AAA tool has been released!
Slack Workspace for Hate Speech Researchers
I created a Slack workspace open to researchers from any instituition (academia and industry) working on the detection of hate speech and other online harms.
Interested in joining? Fill out this form.