Francielle Vargas

I obtained a Ph.D. and M.Sc. in Natural Language Processing, both earned in Computer Science and Computational Mathematics from the University of São Paulo (USP). I also hold a BS in Information Systems and a BA in Linguistics. During my Ph.D., I was a visiting researcher at the University of Southern California (USC) in the USA and an invited researcher and speaker at the Leibniz Institute for the Social Sciences (GESIS) in Germany. I was also awarded the prestigious Google Latin America Research Award (LARA) fellowship, received the Best Thesis Award in Artificial Intelligence, and was finalist for the Computer Society (SBC) Thesis and Dissertation Awards. In addition, I have actively contributed to ACL and AAAI international conferences and workshops, serving as a program committee member and co-organizing International AAAI Conference on Web and Social Media (ICWSM), Workshop on Online Abuse and Harms (WOAH), and Explainable Deep Neural Networks for Responsible AI (DeepXplain), along with shared tasks on hate speech [i] [ii]. My research interests focus on enhancing the interpretability and robustness of Language Models (LMs), while ensuring their alignment with safety, factuality, and fairness. I have contributed to efforts addressing hate speech and misinformation in underrepresented communities, and to the development of methods and benchmark datasets that aim to promote transparency and accountability in AI systems.

📰 Latest News

Sep 24, 2025 | My Ph.D. thesis was selected among the top six best theses in the country in the fields of Multimedia, Hypermedia, and the Web, being recognized as a finalist for the Thesis and Dissertation Award (CTD) at the Brazilian Symposium on Multimedia and the Web (WebMedia).
Set 20, 2025 | Excited to serve on the evaluation committee of the HACKATHON RAIA 2025! The event will provide a unique environment for the development of cutting-edge AI solutions to tackle misinformation, combining hands-on learning, specialized mentorship, and recognition through awards.
Ago 20, 2025 | Our paper has been accepted at EMNLP 2025! The “MFTCXplain: A Multilingual Benchmark Dataset for Evaluating the Moral Reasoning of LLMs through Multi-hop Hate Speech Explanations”, introduces a multilingual benchmark dataset to evaluate how large language models reason morally when handling hate speech, using multi-hop explanations.
Ago 8, 2025 | My Ph.D. thesis has been recognized as the best in the field of Artificial Intelligence in Brazil, receiving the prestigious Maria Carolina Monard Award! Check it out!
July 1, 2025 | Excited to supervise two students at the DeepXplain Summer School on Explainable and Trustworthy AI at the University of Oslo, where we are working on an extension of my PhD research: the Supervised Rational Attention (SRA) method.
May 7, 2025 | My PhD thesis was recognized as one of the best in Computer Science in the country, being selected as a finalist for the Brazilian Computer Society (SBC) Thesis and Dissertation Award (CTD)! Check it out!
Fev 28, 2025 | Our paper "Discourse-aware annotation guideline for low-resource languages" has finally been published in the Natural Language Processing Journal (Cambridge University Press)! By explicitly modeling discourse-level phenomena, our approach not only enhances data quality but also helps to produce more interpretable machine learning models.
Jan 15, 2025 | Honored to be part of the Program Committee for the "Towards a safer web for women: first international workshop on protecting women online" at the web conference 2025", hosted in Sydney, Australia. This workshop will serve as a vital platform for discussing emerging challenges and technological solutions to combat online violence against women.
Dec 5, 2024 | Our IJCNN 2025 Special Session: "Explainable deep neural networks for responsible AI: post-hoc and self-explaining approaches (DeepXplain 2025)" has been accepted! This session will bring innovative discussions on making Deep Neural Networks (DNNs) more transparent and responsible.
Nov 29, 2024 | Our paper has been accepted at COLING 2025! In this paper, we introduce the "HateBRXplain, the first benchmark dataset for explainable hate speech detection in Portuguese".
Nov 12, 2024 | Our paper has been accepted at FEVER @ EMNLP 2024! We introduce the "SELFAR - Sentence-Level Factual Reasoning for Explainable Fact-Checking!"
Sep 06, 2024 | Our paper, "Context-aware and expert data resources for Brazilian Portuguese hate speech detection" has been published online in the Natural Language Processing Journal by Cambridge University Press!
Sep 05, 2024 | Honored to be selected as an organizer for the Workshop on Online Abuse and Harms (WOAH @ ACL 2025) alongside Flor Miriam Plaza del Arco, Agostina Calabrese, Debora Nozza, Christine de Kock and Zeerak Talat. Looking forward to contributing to key discussions on tackling harmful online!
Jun 16, 2024 | Our paper has been accepted at WOAH @ NAACL 2024! We introduced the "HausaHate: The first benchmark dataset for Hate Speech Detection in Hausa!"
Apr 1, 2024 | I am visiting the Computational Social Science Lab at the University of Southern California (USC) in LA under the supervision of Professor Morteza Dehghani. I am excited about exchanging knowledge about disinformation, hate speech, and AI explainability, as well as discussing potential future collaborations!
Fev 16, 2024 | Honored to be a Keynote Speaker at the TechCLAVE 2024 International Conference, organized by the PG Department of Journalism and Communication at DG Vaishnav College in Chennai, India. I will be sharing insights on fine-grained factuality and media bias prediction. I'm looking forward to engaging with global experts on the role of technology in shaping journalism and digital media.
Dec 4, 2023 | Honored to receive the prestigious Google Latin America Research Award (LARA) PhD Fellowship! This recognition highlights my contributions to the field and supports my ongoing research on responsible AI and social impact.
Nov 16, 2023 | Invited to speaker at the Conference on Harmful Online Communication at GESIS – Leibniz Institute for the Social Sciences in Cologne, Germany. My talk provided insights into the methods and benchmarks we developed in Brazil to combat disinformation and hate speech.
Sep 8, 2023 | Three of our papers have been accepted at RANLP 2023! We presented three full papers covering fact-checking, hate speech detection, media bias, and social stereotype bias mitigation.
Sep 4, 2022 | Invited to be a Dataset Co-Chair for ICWSM'23, alongside Kokil Jaidka, Michael Sirivianos, and Siqi Wu!
Jun 20, 2022 | Two of our papers have been accepted at LREC 2022! we introdcued the "HateBR, the first benchmark dataset for hate speech detection in Brazilian Portuguese" and "A Survey on Rhetorical Structure Theory for Online Deception Detection".
Set 1, 2021 | Our paper has been accepted at RANLP 2021! The "Contextual-Lexicon Approach for Abusive Language Detection" introduces a contextualized BoW model that leverages feature saliency to improve the interpretability of hate speech detection.
Jun 8, 2021 | Proud to have contributed to the first Diversity & Accessibility (D&A) initiatives for ICWSM! The report outlines concrete actions and lessons learned to help foster a more inclusive and welcoming conference. Check it out!
April 17, 2020 | Participated in the "Antisemitism on Social Media Datathon and Hackathon", an international project led by Professors Gunther Jikeli and Damir Cavar from NLP reserach group at Indiana University, one of USA's top research universities. Check it out!

Awards, Grants and Scholarships

Research Projects

📰 Latest News