Procheta Sen

Dr. Procheta Sen is a Lecturer in Computer Science at the University of Liverpool's Faculty of Science and Engineering, Department of Computer Science, where she joined in 2022 as part of the Natural Language Processing research group. Her work focuses on developing transparent, fair, and accessible AI language models through explainability research.

She earned her PhD from Dublin City University, Ireland (2021) under supervisor Gareth J.F. Jones, with affiliation at ADAPT Centre Ireland. Prior to Liverpool, she conducted postdoctoral research with Emine Yilmaz in University College London's Web Intelligence Group.

Dr. Sen's research spans three explainability categories: post-hoc methods (feature attribution, counterfactuals), mechanistic interpretability (neural network circuit mapping), and intrinsic interpretability (human-readable model design). She targets diverse end-users including clinicians, legal experts, and laypersons, with applications in bias mitigation and socially responsible AI systems. Her work bridges Natural Language Processing, Machine Learning, and Information Retrieval to address real-world challenges in transparency and equity.

Analysis of her 2023-2025 publications reveals dominant trends in LLM bias analysis, legal document processing, and adaptive conversational systems. Key themes include mechanistic interpretability for bias detection, retrieval-augmented generation for dialogue systems, and multilingual knowledge extraction—demonstrating consistent focus on making AI both technically robust and socially beneficial across domains like law and finance.

Dr. Sen actively advises PhD student Lingfang Li (AAAI 2025 accepted work) and emphasizes compassionate talent development. Her open-source research has been deployed in legal sector applications, and she organizes the annual NLP for Social Good symposium fostering interdisciplinary collaboration for responsible AI.

She leads initiatives including the 2025 virtual NLP for Social Good Symposium and collaborates with institutions like Nokia Bell Labs Cambridge, maintaining active research momentum in transparent AI systems.