Sliced Wasserstein Kernel for Persistence Diagrams

Abstract

Persistence diagrams (PDs) play a key role in topological data analysis (TDA), in which they are routinely used to describe topological properties of complicated shapes. PDs enjoy strong stability properties and have proven their utility in various learning contexts. They do not, however, live in a space naturally endowed with a Hilbert structure and are usually compared with specific distances, such as the bottleneck distance. To incorporate PDs in a learning pipeline, several kernels have been proposed for PDs with a strong emphasis on the stability of the RKHS distance w.r.t. perturbations of the PDs. In this article, we use the Sliced Wasserstein approximation SW of the Wasserstein distance to define a new kernel for PDs, which is not only provably stable but also provably discriminative (depending on the number of points in the PDs) w.r.t. the Wasserstein distance $d_1$ between PDs. We also demonstrate its practicality, by developing an approximation technique to reduce kernel computation time, and show that our proposal compares favorably to existing kernels for PDs on several benchmarks.

Minor modifications

References in corpus (6)

• Statistical topological data analysis using persistence landscapes
• Hierarchical structures of amorphous solids characterized by persistent homology
• Persistence weighted Gaussian kernel for topological data analysis
• Principal Component Analysis of Persistent Homology Rank Functions with case studies of Spatial Point Patterns, Sphere Packing and Colloids
• Kernel method for persistence diagrams via kernel embedding and weight factor
• Comparing persistence diagrams through complex vectors

Cited by in corpus (42)

• On the stability of persistent entropy and new summary functions for TDA
• Topology-Preserving Deep Image Segmentation
• Scalar Field Comparison with Topological Descriptors: Properties and Applications for Scientific Visualization
• PersLay: A Neural Network Layer for Persistence Diagrams and New Graph Topological Signatures
• Persistence paths and signature features in topological data analysis
• The persistence landscape and some of its properties
• On Efficient Optimal Transport: An Analysis of Greedy and Accelerated Mirror Descent Algorithms
• Learning metrics for persistence-based summaries and applications for graph classification
• giotto-tda: A Topological Data Analysis Toolkit for Machine Learning and Data Exploration
• Stochastic Wasserstein Barycenters
• Progressive Wasserstein Barycenters of Persistence Diagrams
• Topological Persistence Machine of Phase Transitions
• Topological Feature Vectors for Chatter Detection in Turning Processes
• Approximating Continuous Functions on Persistence Diagrams Using Template Functions
• Leveraging the Mathematics of Shape for Solar Magnetic Eruption Prediction
• Topological Learning for Brain Networks
• Wasserstein Stability for Persistence Diagrams
• Topological Data Analysis (TDA) for Time Series
• Optimizing persistent homology based functions
• Topology-Aware Segmentation Using Discrete Morse Theory
• Understanding the Topology and the Geometry of the Space of Persistence Diagrams via Optimal Partial Transport
• Localization in the Crowd with Topological Constraints
• Adaptive Partitioning for Template Functions on Persistence Diagrams
• A Progressive Approach to Scalar Field Topology
• A contribution to Optimal Transport on incomparable spaces
• Fuzzy c-Means Clustering for Persistence Diagrams
• Wasserstein Measure Coresets
• Fast Approximation of the Sliced-Wasserstein Distance Using Concentration of Random Projections
• Tree-Sliced Variants of Wasserstein Distances
• Generalized Sliced Distances for Probability Distributions
• Persistence B-Spline Grids: Stable Vector Representation of Persistence Diagrams Based on Data Fitting
• ATOL: Measure Vectorization for Automatic Topologically-Oriented Learning
• Prediction of Carbon Nanostructure Mechanical Properties and Role of Defects Using Machine Learning
• Topological Detection of Trojaned Neural Networks
• TopoTxR: A Topological Biomarker for Predicting Treatment Response in Breast Cancer
• $\zeta$-functions and the topology of superlevel sets of stochastic processes
• Estimation and Quantization of Expected Persistence Diagrams
• Persistence Atlas for Critical Point Variability in Ensembles
• A Domain-Oblivious Approach for Learning Concise Representations of Filtered Topological Spaces for Clustering
• On $C^0$-persistent homology and trees
• Fast Unbalanced Optimal Transport on a Tree
• On Multimarginal Partial Optimal Transport: Equivalent Forms and Computational Complexity