Linus Hamilton is trying to understand what makes neural networks tick.
He graduated with a degree in mathematics from Carnegie Mellon University and is now enjoying MIT as a grad student. His fellowship is funded by a Google Fellowship established through a partnership between the Hertz Foundation and Google.
Linus wants to give researchers guidelines to guess which problems are amenable to deep learning. He also wants to invent tools to help researchers poke inside and understand the neural networks they train.
When he’s not doing math, you can find Linus square-dancing, reading computer-generated poems, or designing games. You can play a bullet-dodging computer game he created. You can also read his game-inspired complexity theory paper ‘Braid is undecidable’ online. Linus grew up in College Park, Maryland.