Changmin Yu
Postdoc in computational neuroscience at CBL, Cambridge
About Me
Hi, I am Changmin Yu (于昌民). I am currently a postdoc in computational neuroscience at Computational and Biological Learning Lab in the engineering department at University of Cambridge, advised by Máté Lengyel. Previously, I completed my PhD in computational neuroscience and machine learning at UCL as a DeepMind scholar, supervised by Neil Burgess and Maneesh Sahani.
My research focuses on normative modelling of hippocampal system with probabilistic inference. I am also interested in deep generative models and reinforcement learning. Previously I worked as research scientists in industries, including Optiver, Meta, Microsoft Research, and Huawei, and was a visiting PhD student in Samuel Gershman's lab. Prior to PhD, I studied mathematics at Imperial College and Cambridge.
In my spare time, I enjoy hiking, playing football and tennis.
Selected Publications/Preprints
International Conference on Machine Learning (ICML), 2026 Spotlight
Accepted in principle at Nature Neuroscience, 2026
International Conference on Learning Representations (ICLR), 2025 Spotlight
Advances in Neural Information Processing Systems (NeurIPS), 2023
Advances in Neural Information Processing Systems (NeurIPS), 2022
International Conference on Learning Representations (ICLR), 2021
News
- May 2026 I have been seleced as a DAAD AINeT fellow.
- April 2026 Our paper "Hierarchical successor representation for robust transfer" has been accepted as a Spotlight at ICML 2026.
- April 2026 Our paper "Hippocampal theta sweeps indicate goal direction during navigation" has been accepted in principle at Nature Neuroscience.
- April 2026 I have been accepted to the Simons Junior Neuroscientist Workshop.
- Jan 2025 Our paper "Discovering temporally compositional neural manifolds with switching infinite GPFA" was accepted as a Spotlight at ICLR 2025.
- Oct 2024 I was awarded a Trainee Professional Development Award at Society for Neuroscience (SfN) 2024.
- Jan 2024 Started postdoc position at CBL, University of Cambridge, working with Máté Lengyel.
- Nov 2023 I successfully defended my PhD thesis "Hippocampus-Inspired Representation Learning for Artificial Agents" with minor corrections, thanks to my examiners Peter Dayan and Andrew Saxe!