Streaming Kernel PCA with $\tilde{O}(\sqrt{n})$ Random Features

Authors: Enayat Ullah, Poorya Mianjy, Teodor Vanislavov Marinov, Raman Arora

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we provide empirical results on a real dataset to support our theoretical results. We perform experiments on the MNIST dataset that consists of 70K samples, partitioned into a training, tuning, and a test set of sizes 20K, 10K, and 40K, respectively.
Researcher Affiliation Academia Department of Computer Science, Johns Hopkins University, Baltimore, MD 21204
Pseudocode Yes Algorithm 1 KPCA with Random Features (Meta Algorithm)
Open Source Code No The paper does not provide an explicit statement about the release of their source code or a link to a code repository for the methodology described.
Open Datasets Yes We perform experiments on the MNIST dataset that consists of 70K samples, partitioned into a training, tuning, and a test set of sizes 20K, 10K, and 40K, respectively.
Dataset Splits Yes We perform experiments on the MNIST dataset that consists of 70K samples, partitioned into a training, tuning, and a test set of sizes 20K, 10K, and 40K, respectively.
Hardware Specification No The paper states that 'Runtime is recorded in a controlled environment; each run executed on identical unloaded compute node.' but does not specify any hardware details such as CPU, GPU models, or memory.
Software Dependencies No The paper does not specify any software dependencies with version numbers.
Experiment Setup Yes In particular, we choose the RBF kernel k(x, x )=exp x x 2/2σ2 with bandwidth parameter σ2 = 50. Number of random features and the size of Nyström approximation are set to 750 and 100, respectively.