Statistical Topological Data Analysis - A Kernel Perspective

Authors: Roland Kwitt, Stefan Huber, Marc Niethammer, Weili Lin, Ulrich Bauer

NeurIPS 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments demonstrate, on a couple of synthetic and real-world data samples, how this universal kernel enables a principled solution to the selected problem of (kernel-based) two-sample hypothesis testing.
Researcher Affiliation Academia Roland Kwitt Department of Computer Science University of Salzburg rkwitt@gmx.at Stefan Huber IST Austria stefan.huber@ist.ac.at Marc Niethammer Department of Computer Science and BRIC UNC Chapel Hill mn@cs.unc.edu Weili Lin Department of Radiology and BRIC UNC Chapel Hill weili_lin@med.unc.edu Ulrich Bauer Department of Mathematics Technische Universität München (TUM) ulrich@bauer.org
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code Yes Source code to reproduce the experiments is available at https://goo.gl/Kou BPT.
Open Datasets Yes The corpus callosum surfaces were obtained from the longitudinal dataset of the OASIS brain database3. 3available online: http://www.oasis-brains.org
Dataset Splits No The paper describes sampling methods and bootstrapping for hypothesis testing, but it does not specify explicit training/validation/test dataset splits (e.g., percentages, counts, or predefined partition files) for model training or evaluation.
Hardware Specification No The paper does not provide specific details about the hardware used for running experiments, such as CPU or GPU models, or memory specifications.
Software Dependencies No The paper mentions 'Dipha' and provides a URL but does not specify its version number or versions for any other software dependencies.
Experiment Setup Yes In all experiments, we use the proposed kernel u-PSS kernel k U σ of Eq. (5) and vary the HKS time ti in 1 = t1 < t2 < < t20 = 10.5; Regarding the u-PSS kernel scale σi, we sweep from 10^-9 = σ1 < < σ10 = 10^1. ... The test statistic under H0 is bootstrapped using B = 5 x 10^4 random permutations.