Learning Adaptive Random Features

Authors: Yanjun Li, Kai Zhang, Jun Wang, Sanjiv Kumar4229-4236

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental This section reports empirical evaluations in numerical kernel matrix approximation and supervised learning tasks.
Researcher Affiliation Collaboration 1University of Illinois at Urbana-Champaign, 2Temple University, 3East China Normal University, 4Google Research
Pseudocode Yes Algorithm 1 Learning Fourier Features
Open Source Code No The paper does not provide an explicit statement or link to its own open-source code for the described methodology.
Open Datasets Yes The benchmark datasets used are listed in Table 1.
Dataset Splits Yes All data samples are split into training/test sets (2 : 1), unless provided in the original data. We tune the parameters via cross validation on training set.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., library names like PyTorch 1.9 or solver versions) needed to replicate the experiments.
Experiment Setup Yes Input data is normalized to have zero mean and unit variance in each dimension, and the Gaussian kernel width 2σ2 is chosen as the dimension d of the input data... We tune the parameters via cross validation on training set. For number of features r = 50, 100, and 200, we run SAMPLE and CLUSTERp with number of landmarks n = r/25, r/5, r, and 5r. ... All the experiments use n = r = 200. For regression, we use ridge regression and report root mean square error (RMSE); for classification, we use ℓ2-regularized SVM and report classification error.