Guillermo Valle Perez

Guillermo Valle Perez
Graduate student
Physics and MMathPhys graduate.
Sysbio DTC
Area of research: Deep learning theory. I focus on generalization theory, and macroscopic properties of large and deep neural networks.
Supervisor: Ard Louis (group page)
Research interests: AI/learning theory, complex systems, algorithmic information theory, statistical physics, soft matter
Other interests: Anything in math, physics, computer science. Making cool stuff (art+technology), AR/VR, philosophy, ...
Personal webpage: http://guillefix.me
Deep learning generalizes because the parameter-function map is biased towards simple functions (to appear in ICLR 2019)