New arXiv paper explores stability-compatible kernels for impulse response estimation
A new paper on kernel-based regularization for linear system identification proposes stability-compatible kernels that go beyond Gaussianity for impulse response estimation, using positive semidefinite kernels to define reproducing kernel Hilbert spaces.
Read the full story at arXiv stat.ML ↗
Timeline · 1 report
- 2026-11-19 05:00 · arXiv stat.ML
Stability-compatible kernels: beyond Gaussianity for impulse response estimation