Online Low-Rank Subspace Clustering by Basis Dictionary Pursuit

Authors: Jie Shen, Ping Li, Huan Xu

ICML 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on synthetic and realistic datasets further substantiate that our algorithm is fast, robust and memory efficient.
Researcher Affiliation Academia Jie Shen JS2007@RUTGERS.EDU Department of Computer Science, Rutgers University, Piscataway, NJ 08854, USA; Ping Li PINGLI@STAT.RUTGERS.EDU Dept. of Statistics & Biostatistics, Dept. of Computer Science, Rutgers University, Piscataway, NJ 08854, USA; Huan Xu ISEXUH@NUS.EDU.SG Department of Industrial and Systems Engineering, National University of Singapore, Singapore
Pseudocode Yes Algorithm 1 Online Low-Rank Subspace Clustering
Open Source Code No The paper does not provide any links to open-source code or explicit statements about code availability.
Open Datasets Yes Datasets. We examine the performance for subspace clustering on 5 realistic databases shown in Table 1, which can be downloaded from the Lib SVM website. For MNIST, We randomly select 20000 samples to form MNIST-20K since we find it time consuming to run the batch methods on the entire database. Table 1. Datasets for subspace clustering. #classes #samples #features Mushrooms 2 8124 112 DNA 3 3186 180 Protein 3 24,387 357 USPS 10 9298 256 MNIST-20K 10 20,000 784
Dataset Splits No The paper mentions datasets used for experiments but does not specify how they were split into training, validation, and test sets, beyond mentioning MNIST-20K was formed by random selection. No percentages or counts for splits are given.
Hardware Specification No The paper states that PCP utilizes a highly optimized C++ toolkit while their algorithms are fully written in Matlab, but it does not provide any specific hardware details (e.g., GPU/CPU models, memory) used for running their experiments.
Software Dependencies No The paper mentions their algorithms are "fully written in Matlab" but does not specify a version number or any other software dependencies with versions.
Experiment Setup Yes Parameters. We set λ1 = 1, λ2 = 1/ p and λ3 = t/p, where t is the iteration counter. These settings are actually used in ORPCA. We follow the default parameter setting for the baselines.