Research map / Learning theory and algorithms / Kernels and regression theory
Robust Persistence Diagrams using Reproducing Kernels
Siddharth Vishwanath, Kenji Fukumizu, Satoshi Kuriki, Bharath K. Sriperumbudur
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace KDE filtration with robust KDE filtration | persistence diagram · Two circles with outliers, noise 20-40% | bottleneck distance to clean diagram | Improved |
| increase outlier noise level from 20% to 40% | robust vs KDE persistence diagram · Two circles with outliers | bottleneck distance advantage | Improved |
| use robust KDE instead of KDE as filter | persistence diagram · Runtime, torus n=1000, resolution 0.10 | runtime: 1.8s -> 3.5s | No clear effect |
| use robust KDE instead of KDE as filter | persistence diagram · Runtime, torus n=1000, resolution 0.04 | runtime: 75.5s -> 76.7s | No clear effect |
| use robust diagram Dgm(f) with W1 instead of persistence image of distance function with L1 | spectral clustering · Six-class 3D shapes, H0 | Rand index: 78.53% -> 95.30% | Improved |
| use robust diagram with W1 instead of persistence image with L1 | spectral clustering · Six-class 3D shapes, H1 | Rand index: 81.89% -> 91.43% | Improved |
| use robust diagram with W1 instead of persistence image with L1 | spectral clustering · Six-class 3D shapes, H2 | Rand index: 80.09% -> 86.33% | Improved |
| use robust diagram with W2 instead of persistence image with L2 | spectral clustering · Six-class 3D shapes, H2 | Rand index: 77.12% -> 73.91% | Hurt |
| use robust diagram with W1 instead of persistence image with L1 on combined max distance | spectral clustering · Six-class 3D shapes, max | Rand index: 82.43% -> 95.72% | Improved |
| use robust KDE persistence images instead of KDE or DTM | linear SVM on persistence images · MPEG7 3-class and 5-class with 15% noise | classification error | Improved |
Closest papers by meaning
- Nonparametric Estimation of Probability Density Functions of Random Persistence Diagrams (JMLR 2019 · similarity 0.91)
- Kernel Method for Persistence Diagrams via Kernel Embedding and Weight Factor (JMLR 2018 · similarity 0.90)
- Persistence Fisher Kernel: A Riemannian Manifold Kernel for Persistence Diagrams (NeurIPS 2018 · similarity 0.89)
- Towards a Persistence Diagram that is Robust to Noise and Varied Densities (ICML 2023 · similarity 0.89)
- Persistence weighted Gaussian kernel for topological data analysis (ICML 2016 · similarity 0.89)
- Persistent Homology for High-dimensional Data Based on Spectral Methods (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
