Chen Sun
I have always been passionately curious about the nature of intelligence, and in particular, about how machines and humans are able to abstract out explanations, deep meaning, and hidden principles from the banalities of everyday life.
My journey began with mathematics. Half a lifetime ago, I was a member of Team Canada at the International Mathematical Olympiad. Though I am no longer a mathematician, it occupies a special place in my life as the first time I encountered abstraction, in what became a lifelong love and obsession.
From there, I switched fields while in Uni, doing neuroscience in Cambridge and then during my PhD at MIT, eventually studying how the brain encodes and consolidates episodic memory. See here for my work here, one of the first ever demonstrations of the coding of events as abstract entities, in the mouse brain. And see here for work between my wife and I on how the brain triages new incoming memories, separating out those worthy for further consolidation.
Presently, I am a Research Scientist at Google DeepMind. My two primary research questions now are how memory consolidation works—both in artificial systems and in brains—and how open-ended processes can perpetually create new ideas and solutions. I believe these two questions are intertwined: robust memory systems enable the continual exploration that leads to genuine novelty, whether in biological organisms or AI models. By focusing on these 2 problems, I hope to understand the shared principles behind learning, remembering, and evolving in complex environments. I am driven by the conviction that understanding these fundamental mechanisms will help us build better AI, inspire deeper insights into how our brains work, and ultimately expand the science of what is possible.
