Streaming PCA for Markovian Data

Authors: Syamantak Kumar, Purnamrita Sarkar

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In Section 6, the paper presents 'Experimental Validation' with figures comparing Oja's algorithm performance.
Researcher Affiliation Academia Syamantak Kumar1 Purnamrita Sarkar2 1Department of Computer Science, UT Austin 2Department of Statistics and Data Sciences, UT Austin syamantak@utexas.edu, purna.sarkar@austin.utexas.edu
Pseudocode No The paper provides the update rule for Oja's algorithm as an equation (Eq. 1) but does not include a structured pseudocode block or algorithm listing.
Open Source Code No The paper does not provide any statement about making its source code available or include a link to a code repository.
Open Datasets No The paper uses synthetic data generated internally: 'Each state s Ωis associated with D(s) := Bernoulli(ps) distribution. We set d = 1000 and select ps U (0, 0.05) at the start of each random run.' It does not refer to a publicly available dataset with concrete access information.
Dataset Splits No The paper describes generating data for experiments but does not specify any training, validation, or test dataset splits. The data is generated 'on the fly' for each random run.
Hardware Specification No The paper describes the experimental setup but does not provide any specific details about the hardware (e.g., GPU, CPU models) used to run the experiments.
Software Dependencies No The paper does not specify any software dependencies with version numbers (e.g., programming languages, libraries, frameworks) used for the experiments.
Experiment Setup Yes The step sizes for Oja s algorithm are set as ηi = α (β+i)(λ1 λ2) for α = 5, β = 5 1 |λ2(P )|. For the downsampled variant, every 10th data-point is considered, and β is accordingly divided by 10.