WorksMathematics
Smoothing Bias and Direction Recovery at Boundary Minima of the Nuclear Norm along Matrix Lines
Abstract
We study optimal matrix directions associated with minimizing , the nuclear norm along a fixed affine line of real matrices with scalar multiplier . At a rank-deficient minimum, the compact polar factor can violate the tangency constraint even at the minimizing multiplier. Spectral smoothing gives a continuous scalar equation for tangency. We quantify its bias at a unique boundary minimum, where exactly one one-sided derivative vanishes. An orthogonal block decomposition shows that pseudo-Huber smoothing with scale shifts the multiplier by exact order as . The Frobenius error relative to the minimum-Frobenius-norm optimal direction has exact order when some singular values vanish at exactly second order in the multiplier displacement, and otherwise. A sharp rank bound rules out the slower rate when has rank at most two. Huber smoothing gives an error of exact order for the same limiting direction. Convergent expansions in powers of also yield a three-scale extrapolation with error after restoring feasibility.