Inverse M-Kernels for Linear Universal Approximators of Non-Negative Functions

Authors: Hideaki Kim

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We confirm the effectiveness of our results by experiments on the problems of non-negativity-constrained regression, density estimation, and intensity estimation.
Researcher Affiliation Industry Hideaki Kim NTT Corporation hideaki.kin@ntt.com
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code Yes Code and data to reproduce the results are available at https://github.com/Hid Kim/IM-Kernel.
Open Datasets Yes Code and data to reproduce the results are available at https://github.com/Hid Kim/IM-Kernel.
Dataset Splits Yes The hyper-parameters for each model were optimized through three-fold cross validation on a grid
Hardware Specification Yes A Mac Book Pro with 12-core CPU (Apple M2 Max) was used.
Software Dependencies Yes We implemented all compared models by using Python-3.10.8 (Sci Py-1.11, fnnls-1.0 (MIT License))1.
Experiment Setup Yes The hyper-parameters for each model were optimized through three-fold cross validation on a grid: for NCM, QNM, and IMK, the grid is (τ, r) C C for C = {0.1, 0.2, 0.5, 1, 2, 5, 10}; for SNF, the number of components for Gaussian mixture measure dµ( ) was selected from {1, 2, 3}.