Universality Laws for High-Dimensional Learning with Random Features

September 16, 2020 Β· Declared Dead Β· πŸ› IEEE Transactions on Information Theory

πŸ‘» CAUSE OF DEATH: Ghosted
No code link whatsoever

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Hong Hu, Yue M. Lu arXiv ID 2009.07669 Category cs.IT: Information Theory Citations 157 Venue IEEE Transactions on Information Theory Last Checked 4 months ago
Abstract
We prove a universality theorem for learning with random features. Our result shows that, in terms of training and generalization errors, a random feature model with a nonlinear activation function is asymptotically equivalent to a surrogate linear Gaussian model with a matching covariance matrix. This settles a so-called Gaussian equivalence conjecture based on which several recent papers develop their results. Our method for proving the universality theorem builds on the classical Lindeberg approach. Major ingredients of the proof include a leave-one-out analysis for the optimization problem associated with the training process and a central limit theorem, obtained via Stein's method, for weakly correlated random variables.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

πŸ“œ Similar Papers

In the same crypt β€” Information Theory

Died the same way β€” πŸ‘» Ghosted