Yuechen Jiang

PhD Researcher in Human-Centred AI, Collective Intelligence, and Digital Humanities

I am a PhD researcher at the University of Manchester and a member of the National Centre for Text Mining (NaCTeM). My doctoral research is jointly supervised by Dr. Konstantinos Arvanitis and Professor Sophia Ananiadou, in collaboration with the Getty Conservation Institute, where my industry supervisor is Stavroula Golfomitsou, Head of Collections. This research is supported by a UKRI EPSRC Industrial Doctoral Landscape Award. Before beginning my PhD in January 2026, I worked as a Research Assistant at Stevens Institute of Technology from August 2023, working with the Cognition Lab led by Professor Jordan W. Suchow on large language model agents, memory architectures, and evaluation benchmarks, contributing to projects including FinMem, FinCon, FinBen, and InvestorBench. I received two master's degrees from Stevens Institute of Technology: an M.S. in Machine Learning and an M.S. in Financial Engineering, with concentrations in Machine Learning in Finance and Financial Statistics.

Collective intelligence rarely comes from agreement; it comes from how disagreement gets resolved, when experts with different training, roles, and cultural assumptions manage to converge on a decision without erasing what made their perspectives different in the first place. My research is primarily centred on whether AI systems can participate in this process as genuine epistemic partners, rather than as tools that simply aggregate or defer to human judgment. How can AI agents maintain explicit belief states that evolve through evidence and interaction, rather than treating reasoning as a sequence of independent predictions?

Museum conservation brings these questions into sharp focus. Whether assessing authenticity, planning conservation treatments, or evaluating cultural significance, experts must constantly balance incomplete evidence, conflicting interpretations, and evolving knowledge. Consensus is therefore not a fixed destination but an ongoing process of negotiating evidence and revising beliefs. Understanding this process offers a unique opportunity to rethink how AI systems should reason alongside human experts.