NLP for Socially Grounded Research
By Patricia Chiril

Patricia Chiril will give a talk on NLP for Socially Grounded Research

Abstract

In this talk, I will present an overview of my research career, which can be situated at the intersection of language, society, and technology. Since my PhD, I have applied Natural Language Processing (NLP) techniques to newly created corpora that I have compiled and released to the research community, with the goal of addressing socially relevant issues. Over the years, I have developed a body of work oriented towards exploring and modelling meaning construal in both reporting (journalistic, academic, etc.) and sharing (social media platforms) contexts. To this end, I have worked with diverse natural language data with a view to investigating hate speech detection and uncovering gender representational bias, as well as examining how language is used to frame sustainability and environmental discourse in contexts engaging various stakeholder perspectives. Following technological advances, my research spans a range of architectures, from early neural models (preceding BERT) to the most recent ones. I have explored questions of generalization and fairness, which remain central even with the advent of LLMs. Drawing on examples from my scientific productions, I will reflect on how NLP models capabilities can be leveraged and where their limitations can be overcome to support more socially grounded research. I will also briefly mention some of the international collaborations and interdisciplinary work I am at present engaged in.

Biography

Patricia Chiril joined the Data Science Institute at the University of Chicago as a Postdoctoral Scholar in 2022, after completing her PhD at the University of Toulouse. Her research focuses on Natural Language Processing (NLP), with particular interest in detecting hate speech and algorithmic bias in multilingual social media data, as well as applying NLP to socially grounded domains such as environmental discourse. Her dissertation on sexism detection in French social media was awarded the Prix de Thèse ATALA in 2022, an honorary distinction recognizing outstanding PhD research in NLP. Outside academia, she enjoys skiing, hiking, and photography.