Alexander Ku

I'm a PhD student in psychology and neuroscience at Princeton University, where I'm advised by Tom Griffiths and Jon Cohen. I'm also a part-time research scientist at Google DeepMind.

My research broadly focuses on the dynamics of memory, learning and abstraction in neural networks. Recently, I've been particularly interested in the interaction between gradient-based learning and in-context inference in transformers, due to its resemblance to neocortical and hippocampal learning in the brain. I approach these topics through the lens of rational (Bayesian) analysis, meta-learning, and information theory.

Email / Google Scholar / GitHub / Twitter / CV