On the Saturation Effects of Spectral Algorithms in Large Dimensions
This work provides theoretical insights into the limitations of spectral algorithms in high-dimensional machine learning, which is incremental as it extends known saturation effects to new settings.
The paper investigates saturation effects in spectral algorithms like kernel ridge regression and gradient descent in high-dimensional settings, establishing improved minimax lower bounds and exact convergence rates that reveal periodic plateau behavior and a polynomial approximation barrier. It shows that saturation occurs under different source condition thresholds in large versus fixed dimensions.
The saturation effects, which originally refer to the fact that kernel ridge regression (KRR) fails to achieve the information-theoretical lower bound when the regression function is over-smooth, have been observed for almost 20 years and were rigorously proved recently for kernel ridge regression and some other spectral algorithms over a fixed dimensional domain. The main focus of this paper is to explore the saturation effects for a large class of spectral algorithms (including the KRR, gradient descent, etc.) in large dimensional settings where $n \asymp d^γ$. More precisely, we first propose an improved minimax lower bound for the kernel regression problem in large dimensional settings and show that the gradient flow with early stopping strategy will result in an estimator achieving this lower bound (up to a logarithmic factor). Similar to the results in KRR, we can further determine the exact convergence rates (both upper and lower bounds) of a large class of (optimal tuned) spectral algorithms with different qualification $τ$'s. In particular, we find that these exact rate curves (varying along $γ$) exhibit the periodic plateau behavior and the polynomial approximation barrier. Consequently, we can fully depict the saturation effects of the spectral algorithms and reveal a new phenomenon in large dimensional settings (i.e., the saturation effect occurs in large dimensional setting as long as the source condition $s>τ$ while it occurs in fixed dimensional setting as long as $s>2τ$).