Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].

Square Root Principal Component Pursuit: Tuning-Free Noisy Robust Matrix Recovery

Authors: Junhui Zhang, Jingkai Yan, John Wright

NeurIPS 2021 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We validate the effectiveness of our new method through experiments on simulated and real datasets. Our simulations corroborate the claim that a universal choice of the regularization parameter yields near optimal performance across a range of noise levels, indicating that the proposed method outperforms the (somewhat loose) bound proved here.
Researcher Affiliation Academia Junhui Zhang Department of Applied Physics and Applied Math Columbia University New York, NY 10027 EMAIL Jingkai Yan Department of Electrical Engineering Columbia University New York, NY 10027 EMAIL John Wright Department of Electrical Engineering Columbia University New York, NY 10027 EMAIL
Pseudocode Yes Algorithm 1 Algorithm for PCP
Open Source Code No The paper does not provide any concrete access to source code for the methodology described.
Open Datasets Yes We use the hall dataset in [30]... We use the Dark Raw Video (DRV) dataset in [32] (under MIT License)...
Dataset Splits No The paper does not specify exact training, validation, or test dataset splits (e.g., percentages or absolute counts).
Hardware Specification Yes We run the experiments on a laptop with 2.3 GHz Dual-Core Intel Core i5
Software Dependencies No The paper mentions using 'Matlab' and functions like 'rgb2gray()' and 'imadjustn()', but no specific version numbers are provided for Matlab or any other software dependencies.
Experiment Setup Yes For the added noise, we choose σ {0, 30, 60, 90, 120}... We take λ = 1/ n1, µroot = p n2/2 and µstable = 1 σ( n1+ n2)... set the maximal iteration of ADMM to be 5000.