Dr Zhongtian Sun

Dr Zhongtian Sun is a Lecturer in Artificial Intelligence and Machine Learning in the School of Computing at the University of Kent. He is also a Visiting Fellow in the Department of Computer Science and Technology at the University of Cambridge and a Mila AI Policy Fellow.

His research programme focuses on trustworthy artificial intelligence and representation learning, with particular interests in geometric and topological machine learning, graph and hypergraph learning, causal inference, AI safety, and reliable learning and reasoning. His work develops new methods for understanding and improving the reliability, robustness and interpretability of modern AI systems.

His research outputs span major international AI and machine-learning conferences and journals, including NeurIPS, ICML, ACL, The Web Conference, RecSys, ICAIF, AIED, IEEE Transactions on Neural Networks and Learning Systems (TNNLS), AI Open and IJCNN. Applications of this research span healthcare, financial systems, education and other settings in which reliability, interpretability and accountability are important.

Dr Sun holds a PhD in Computer Science from Durham University. Before joining Kent, he conducted research at the University of Cambridge on machine learning for neuroscience and at the University of Oxford on large language models, knowledge graphs and reasoning. He previously worked in asset management, with experience in financial data analysis, quantitative research and investment-related applications of machine learning.

Alongside his academic research, Dr Sun contributes to international discussions on trustworthy and frontier AI. In 2026, he was selected as an expert to participate in the European Commission AI Office Expert Forum on Frontier AI and contributed expert input to the report Enhancing Competitiveness, Sovereignty and Security of the European Union in Frontier AI. He is also a Mila AI Policy Fellow and works on AI safety and system-level risks in agentic AI.  

Dr Sun's research focuses on trustworthy AI, representation learning, and the foundations of learning, reasoning and adaptation in intelligent systems. He investigates how representations and capabilities emerge and evolve, how they support generalisation and reasoning, and how intelligent systems can interact and make decisions reliably. His work draws on geometric, topological and structured machine learning, with broader interests in foundation models and autonomous AI systems.

Trustworthy AI and AI Safety

Research on trustworthy and reliable AI systems, including robustness, generalisation, interpretability, uncertainty, AI safety, reliable reasoning, and the evaluation of foundation models and emerging intelligent systems.

Geometric, Topological and Structured Representation Learning

Developing new approaches to representation learning using geometry, topology, graphs, hypergraphs, sheaf theory and causal structure. Related interests include geometric deep learning, topological machine learning, graph and hypergraph neural networks, causal representation learning, and the structure and dynamics of learned representations.

Reliable AI in High-Stakes and Human-Facing Systems

Investigating reasoning, adaptation, interaction and decision-making in intelligent systems, including large language models, world models, neuro-symbolic reasoning, agentic and multi-agent AI, and human–AI collaboration. Applications include healthcare, finance, education and other settings where reliability and accountability are important.  

Module Co-Lead responsibilities include module and curriculum design, coordination of teaching, and assessment design and delivery.

• COMP7019 and COMP6004 — Machine Learning Algorithms

Current Research Opportunity — Paid Research Assistant (6 - 12 Months)

I am currently recruiting a Research Assistant for a six-month paid applied AI project in collaboration with a healthcare-sector organisation. Candidates with experience in Python, LLMs, AI agents, information retrieval or RAG are encouraged to send me emails with a CV and a brief summary of relevant experience. 

Prospective PhD and MSc by Research Students

Dr Sun welcomes enquiries from prospective PhD and MSc by Research students interested in machine learning and artificial intelligence. His research interests span trustworthy AI, geometric and topological machine learning, structured representation learning, reasoning and adaptation, foundation models, world models, and autonomous AI systems.

He welcomes a broad range of theoretical, methodological and interdisciplinary research ideas related to these areas, including topics that extend beyond his current projects.

Interested students are encouraged to contact Dr Sun by email with a CV and a brief description of their background and research interests. An initial research idea is welcome, but a fully developed proposal is not required for an initial discussion.

Dr Sun supervises doctoral researchers at the University of Kent and co-supervises doctoral researchers through research collaborations with the University of Cambridge and the University of Oxford. He also contributes to external PhD supervision and welcomes opportunities for external co-supervision where projects align closely with his research programme.

Research leadership and academic service

• Competition Track Chair, ACM International Conference on AI in Finance (ICAIF 2026)
• General Chair, WIVACE 2026
• Area Chair, ICASSP 2026
• Area Chair, ICLR XAI4Science
• Area Chair, AAAI XAI4Science
• TKDE Special Track Organiser
• Senior PC Member and Regular Reviewer for major AI and machine-learning conferences including ICLR, ICML, NeurIPS, AAAI, ACL, ECAI, IJCNN, AIED, The Web Conference (WWW) and RecSys
• Reviewer for international journals including PNAS, Expert Systems with Applications, and IEEE Transactions on Knowledge and Data Engineering (TKDE)
• Association for Computing Machinery (ACM) Member
• Member, IEEE Task Force on Learning for Graphs

International appointments and policy engagement

• Visiting Fellow, Department of Computer Science and Technology, University of Cambridge
• Mila — Quebec AI Institute, AI Policy Fellow
• Selected Expert, European Commission AI Office Expert Forum on Frontier AI, 2026
• Contributor to the European Commission AI Office report Enhancing Competitiveness, Sovereignty and Security of the European Union in Frontier AI

European Commission AI Office report

Published Written Evidence, UK House of Lords Financial Services Regulation Committee (2026):

 Regulatory sandboxes as learning institutions: bounded assurance and reusable evidence in financial AI

Research translation and industry collaboration

• Academic Lead on an Innovate UK Knowledge Transfer Partnership (KTP)
• Research collaborations with major financial institutions on trustworthy AI, responsible AI and AI in high-stakes financial environments
• Collaborative projects with healthcare and industry partners on AI safety, trustworthy AI and reliable AI systems

Invited talks and public engagement

Invited Speaker — Mila AI Policy Conference, Montreal, Canada

Talk: Auditing System-Level Risks in Agentic AI

YouTube: LINK

Podcast Guest — Integrated Cancer Medicine: Research in Focus, University of Cambridge

Invited to discuss explainable multimodal AI for early cancer detection, including Hereditary Diffuse Gastric Cancer (HDGC) and Signet Ring Cell Carcinoma (SRCC), as well as multi-agent decision support grounded in clinical reasoning.

Spotify: LINK

Podcast Guest — TIRIgogy ConnectED, 26th International Conference on Artificial Intelligence in Education

Invited to discuss Teaching, Tracing, and the AI-native Learner, drawing on SPAR-GNN research and teaching practice, including identifying and supporting at-risk learners, selective AI feedback, and teaching in the generative AI era.

Spotify: LINK

Publications

Dr Sun's research outputs span trustworthy AI, representation learning, geometric and topological machine learning, causal learning, graph and hypergraph learning, large language models and reliable AI systems.

His work appears across major international AI and machine-learning conferences and journals, including NeurIPS, ICML, ACL, The Web Conference (WWW), RecSys, ICAIF, AIED, IEEE Transactions on Neural Networks and Learning Systems (TNNLS), AI Open and IJCNN.  

Additional Links

• University of Cambridge Profile
• Mila AI Policy Fellowship