Reproduce “Tensor Decomposition Meets RKHS: Efficient Algorithms for Smooth and Misaligned Data” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Brett W. Larsen, Tamara G. Kolda, Anru R. Zhang, Alex H. Williams
CP-HiFi is introduced as a tensor decomposition method that extends Canonical Polyadic decomposition to 'quasitensors' with continuous modes by integrating Reproducing Kernel Hilbert Spaces. Numerical experiments show it reconstructs underlying continuous functions more accurately and robustly than standard CP, particularly with misaligned or sparse data.
Key insights
- Integrating RKHS with CP decomposition provides a principled and computationally efficient mechanism to inherently enforce smoothness constraints on continuous tensor modes.
- The Representer Theorem is crucial for reducing the complexity of functional optimization in RKHS to a solvable finite-dimensional matrix problem.
- Modeling continuous modes directly allows for natural accommodation of misaligned and irregularly sampled data, overcoming a major limitation of standard tensor decomposition methods.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.