About me

I'm a physicist interested in building a general theory of scales of description. As the brain extracts macroscopic descriptions by its nature (calling a bunch of pixels "cat"), I'm currently doing a Ph.D. in computational neuroscience. Using mathematical and computational tools, I study the computation of extracting a macroscopic description given some data. We found that a system that can extract macroscopic descriptions can successfully solve intelligence tests without any pretraining, which is unique in machine learning. Therefore, we are currently working on publishing these results, relating the ability to extract a macroscopic description without pretraining to human learning-independent abstract reasoning. While we attempt to publish these results (which takes too much time), I'm working on several approaches for a general theory of scales. I'll start publishing this work once I finish the cognitive science work. In the meantime, you can read the Blog. Stay tuned.

Publications