English

Spectral Non-Convex Optimization for Dimension Reduction with Hilbert-Schmidt Independence Criterion

Machine Learning 2019-09-12 v1 Machine Learning Applications Computation

Abstract

The Hilbert Schmidt Independence Criterion (HSIC) is a kernel dependence measure that has applications in various aspects of machine learning. Conveniently, the objectives of different dimensionality reduction applications using HSIC often reduce to the same optimization problem. However, the nonconvexity of the objective function arising from non-linear kernels poses a serious challenge to optimization efficiency and limits the potential of HSIC-based formulations. As a result, only linear kernels have been computationally tractable in practice. This paper proposes a spectral-based optimization algorithm that extends beyond the linear kernel. The algorithm identifies a family of suitable kernels and provides the first and second-order local guarantees when a fixed point is reached. Furthermore, we propose a principled initialization strategy, thereby removing the need to repeat the algorithm at random initialization points. Compared to state-of-the-art optimization algorithms, our empirical results on real data show a run-time improvement by as much as a factor of 10510^5 while consistently achieving lower cost and classification/clustering errors. The implementation source code is publicly available on https://github.com/endsley.

Keywords

Cite

@article{arxiv.1909.05097,
  title  = {Spectral Non-Convex Optimization for Dimension Reduction with Hilbert-Schmidt Independence Criterion},
  author = {Chieh Wu and Jared Miller and Yale Chang and Mario Sznaier and Jennifer Dy},
  journal= {arXiv preprint arXiv:1909.05097},
  year   = {2019}
}

Comments

arXiv admin note: substantial text overlap with arXiv:1909.03093

R2 v1 2026-06-23T11:12:23.426Z