Conditioning of Random Feature Matrices: Double Descent and Generalization Error
This provides theoretical insights into the double descent phenomenon in machine learning, which is foundational for understanding generalization in overparameterized models.
The paper tackles the conditioning of random feature matrices, showing that they are well-conditioned when the complexity ratio scales logarithmically, and proves that regression risk exhibits the double descent phenomenon due to this conditioning behavior, with risk decreasing as data and neurons increase, matching optimal scaling in the literature.
We provide (high probability) bounds on the condition number of random feature matrices. In particular, we show that if the complexity ratio $\frac{N}{m}$ where $N$ is the number of neurons and $m$ is the number of data samples scales like $\log^{-1}(N)$ or $\log(m)$, then the random feature matrix is well-conditioned. This result holds without the need of regularization and relies on establishing various concentration bounds between dependent components of the random feature matrix. Additionally, we derive bounds on the restricted isometry constant of the random feature matrix. We prove that the risk associated with regression problems using a random feature matrix exhibits the double descent phenomenon and that this is an effect of the double descent behavior of the condition number. The risk bounds include the underparameterized setting using the least squares problem and the overparameterized setting where using either the minimum norm interpolation problem or a sparse regression problem. For the least squares or sparse regression cases, we show that the risk decreases as $m$ and $N$ increase, even in the presence of bounded or random noise. The risk bound matches the optimal scaling in the literature and the constants in our results are explicit and independent of the dimension of the data.