Li Nanbo 
Postdoctoral Fellow · Center of Excellence GenAI · KAUST
I am a Postdoctoral Fellow with Jürgen Schmidhuber at the KAUST Center of Excellence for Generative AI. I received my PhD from the School of Informatics at the University of Edinburgh, supervised by Robert Fisher and Chris Williams.
My research spans several areas of AI, particularly sequence modelling, self-improving systems, and world models, with structural computation as the unifying theme.
Recent updates
-
Excited to release our survey on Next-Gen Architectures of LLMs as a preprint!
-
Excited to release NoteBank, our new work on LLM long-context capability: unlocking million-token memory, almost for free!
-
Self-Improvements in Modern Agentic Systems: A Survey was released as a preprint on arXiv.
-
Our Huxley-Gödel Machine paper was accepted at ICLR 2026 as an Oral presentation.
-
SlotSPE, our work on structural prognostic event modelling, was accepted at ICLR 2026.
Earlier updates
-
Awarded a Google Cloud Academic Research Grant.
-
Led a successful application for GenAI CoE & Google Gift Funding.
-
FACTS, our factored state-space framework for world modelling, was accepted at ICLR 2025 as a poster.
-
FACTS was released as a preprint on arXiv.
-
Back to AI research! Started a postdoc at KAUST GenAI CoE, working with Jürgen Schmidhuber.
-
Took a career break to explore the meaning of life. Traveled, learned, and reflected.
-
Received my PhD from the University of Edinburgh.
