Scaling Laws in Linear Regression: Compute, Parameters, and Data

From the perspective of statistical learning theory, (1) is rather intriguing. Moreover, they do not provide instance-wise matching lower bounds to verify the tightness of the upper bounds.