معرفی
Andrew Parry is a PhD Student and Tutor in the School of Computing Science at the University of Glasgow. His research focuses on advancing information retrieval (IR) systems through machine learning and large language models (LLMs), with particular expertise in ranking models, interpretability frameworks, and evaluation methodologies. He has contributed to frameworks like PyTerrier and SimIIR, and his work bridges theoretical IR advancements with practical applications in generative search and relevance assessment.
Education: Pursuing PhD in Computing Science at the University of Glasgow.
Research Interests: Parry explores mechanistic interpretability of IR systems, the role of LLMs in relevance assessment, and the ethical implications of neural retrieval models. His work emphasizes transparency in search algorithms and the development of robust evaluation techniques for large-scale systems.
Key Publications: He has published at top venues including SIGIR, WSDM, and ECIR, with notable contributions on declarative RAG pipelines, LLM-based relevance assessment discrepancies, and corpus subsampling strategies. His 2025 work on MechIR proposes a novel interpretability framework, while recent collaborations address contamination in ranking distillation and interactive search simulations.
Advising & Grants: As a Tutor, Parry supports teaching and research supervision within the School of Computing Science. His research is aligned with broader efforts in the Glasgow Information Retrieval Group, though specific grant details are not disclosed in available texts.
Labs/Teams: Active member of the University of Glasgow's Terrier Team, contributing to TREC evaluations and collaborative workshops like ECIR's Collab-a-thon.
