Timothy Baldwin

Timothy Baldwin is a Professor at the University of Melbourne, School of Computing and Information Systems, with additional affiliation at Mohamed bin Zayed University of Artificial Intelligence in UAE. His research spans natural language processing, large language models, and multilingual AI systems.

His research interests focus on the safety, reliability, and ethical aspects of large language models. He investigates bias evaluation and debiasing techniques, uncertainty quantification methods, fact-checking systems, and multilingual model safety. His work addresses critical challenges in making AI systems more transparent, reliable, and culturally aware, with particular attention to low-resource languages and cross-cultural differences.

Baldwin's recent publications demonstrate a strong focus on evaluating and improving the safety of language models across diverse linguistic contexts, developing tools for fact verification, and understanding the internal mechanisms of large language models. His research shows increasing emphasis on practical applications with real-world impact, particularly in multilingual settings and safety-critical domains.

His scientific contributions include foundational work on multilingual NLP, bias mitigation techniques, and frameworks for evaluating LLM safety across different cultural contexts. His research has been published in top-tier venues including ACL, NAACL, EMNLP, and ICLR.

Baldwin actively mentors students and junior researchers, with frequent collaborations with Haonan Li, Xudong Han, and Fajri Koto, among others. His research group appears to focus on practical applications of NLP with strong ethical considerations, particularly regarding model safety and cultural sensitivity.