Searcharxiv⌕ Search

arXiv subjects

Joseph Gabet

Publications and source records attributed to Joseph Gabet.

2 recordsLinked to original sources

Spectral Conditioning of Weighted Multivariate Vandermonde Operators via Radial Beurling--Selberg Approximation

We establish explicit eigenvalue bounds for Gram matrices of weighted multivariate Vandermonde operators with separated nodes on a torus. The squared magnitudes of the weights are samples of a nonnegative radial spectral energy density whose profile has bounded variation, finite energy, and finite weighted variation cost; compact support is not required. The estimate holds in every dimension and depends on the periodic Euclidean separation and the weighted variation integral. For compactly supported profiles, our analysis combines a signed layer decomposition with the ball approximants of Holt and Vaaler; the noncompactly supported case follows by truncation. Combining the resulting radial envelopes with a bandlimited sampling identity based on Poisson summation yields the spectral bounds by eliminating off-diagonal interactions and nonzero diagonal aliases. The result covers nonmonotone profiles and discontinuous radial low-pass filters, and it is complemented by a global elementary upper bound on the Holt--Vaaler cost that exhibits two radial regimes.

math.NA↗

Global Convergence of ESPRIT with Preconditioned First-Order Methods for Spike Deconvolution

Spike deconvolution is the problem of recovering point sources from their convolution with a known point spread function, playing a fundamental role in many sensing and imaging applications. This paper proposes a novel approach combining ESPRIT with Preconditioned Gradient Descent (PGD) to estimate the amplitudes and locations of the point sources by a non-linear least squares. The preconditioning matrices are adaptively designed to account for variations in the learning process, ensuring a proven super-linear convergence rate. We provide local convergence guarantees for PGD and performance analysis of ESPRIT reconstruction, leading to global convergence guarantees for our method in one-dimensional settings with multiple snapshots, demonstrating its robustness and effectiveness. Numerical simulations corroborate the performance of the proposed approach for spike deconvolution.

eess.SP↗