Debora Nozza

Debora Nozza is an Assistant Professor in Computing Sciences at Bocconi University, where she leads research at the intersection of Natural Language Processing, ethics, and social impact. Her work focuses on detecting and countering hate speech, understanding algorithmic bias, and examining how people use large language models in everyday contexts.

Education:

• PhD in Computer Science, 2018, University of Milano Bicocca

Dr. Nozza's research centers on ethical challenges in Natural Language Processing, with particular emphasis on hate speech detection across multilingual contexts and bias mitigation in language technologies. Her work bridges technical NLP methods with social science perspectives, examining how language models interact with societal structures and cultural contexts. She has made significant contributions to understanding gender bias in machine translation and developing methods for more equitable language technologies.

Her recent publications reveal a strong focus on practical applications of NLP for social good, with an increasing emphasis on evaluating and mitigating bias in multimodal systems and large language models. Her research spans technical methodology development, dataset creation for under-resourced languages, and critical analysis of evaluation metrics in bias research.

Scientific Awards:

• €1.5 million ERC Starting Grant (2023) for research on personalized and subjective approaches to Natural Language Processing
• €120,000 grant from Fondazione Cariplo for the MONICA project on monitoring Italian measures in response to COVID-19

Dr. Nozza actively contributes to the NLP community through her leadership in the MilaNLP group. She has secured significant research funding that supports multiple researchers working on NLP for social impact. Her work has practical implications for social media platforms, policymakers, and organizations developing language technologies.

She co-leads the MilaNLP research group at Bocconi University, which focuses on Natural Language Processing with social impact. The group has been active in major NLP conferences, organizing workshops on online abuse and gender bias, and contributing to shared tasks that advance the field's understanding of hate speech and bias in multilingual contexts.