On the Computational Power of Online Gradient Descent
Vaggos Chatziafratis, Tim Roughgarden, Joshua R. Wang · arXiv (Cornell University) · 2018
We prove that the evolution of weight vectors in online gradient descent can encode arbitrary polynomial-space computations, even in very simple learning settings. Our results imply that, under weak complexity-theoretic assumptions, it is impossible to reason efficiently about the fine-grained behavior of online gradient descent.