Interaction Measures, Partition Lattices and Kernel Tests for High-Order Interactions
Authors: Zhaolu Liu, Robert Peach, Pedro A.M Mediano, Mauricio Barahona
NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We illustrate our results numerically with validations on synthetic data, and through an application to neuroimaging data. 6 Experiments We first validate our family of d-order interaction tests (in particular, d = 5) on synthetic datasets with ground truth interactions, and then investigate their application to a neuroimaging dataset. |
| Researcher Affiliation | Academia | 1Department of Mathematics, Imperial College London, United Kingdom 2Department of Neurology, University Hospital Würzburg, Germany 3Department of Brain Sciences, Imperial College London, United Kingdom 4Department of Computing, Imperial College London, United Kingdom |
| Pseudocode | Yes | Algorithm 1 Permutation test for the interaction measures |
| Open Source Code | Yes | This code for performing the interaction tests and the synthetic experiments is provided in this anonymous Github repository https://github.com/barahona-research-group/ streitberg-interaction.git. |
| Open Datasets | Yes | The dataset consists of resting-state f MRI data from 50 unrelated subjects part of the Human Connectome Project [50, 51]. |
| Dataset Splits | No | Unless stated otherwise, the significance level is set to α = 0.05, sample size to n = 80, and we use Gaussian kernels with the median heuristic as the bandwidth. The paper discusses generating data for synthetic experiments and using fMRI data, but does not provide explicit train/validation/test dataset splits beyond this setup information. |
| Hardware Specification | Yes | All experiments carried out on a 2015 i Mac with 4 GHz Quad-Core Intel Core i7 processor and 32 GB 1867 MHz DDR3 memory. |
| Software Dependencies | Yes | All preprocessing steps were performed using the CONN toolbox (https://www. nitrc.org/projects/conn/), version 17f58. |
| Experiment Setup | Yes | Unless stated otherwise, the significance level is set to α = 0.05, sample size to n = 80, and we use Gaussian kernels with the median heuristic as the bandwidth. |