Self-Paced Learning for Matrix Factorization

Authors: Qian Zhao, Deyu Meng, Lu Jiang, Qi Xie, Zongben Xu, Alexander Hauptmann

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate the performance of the proposed SPMF approach...on synthetic, structure from motion and background subtraction data.
Researcher Affiliation Academia 1School of Mathematics and Statistics, Xi an Jiaotong University 2School of Computer Science, Carnegie Mellon University
Pseudocode Yes Algorithm 1 Self-paced matrix factorization algorithm
Open Source Code No The paper mentions using 'publicly available codes from the authors websites' for competing methods, but does not state that its own code is publicly available or provide a link.
Open Datasets Yes For rigid SFM, we employ the Dinosaur sequence1 which contains 319 feature points tracked over 36 views... 1http://www.robots.ox.ac.uk/ abm/. For nonrigid SFM, we use the Giraffe sequence2, which includes 166 feature points tracked over 120 frames. 2http://www.robots.ox.ac.uk/ abm/. Four video sequences provided by Li et al. (2004)4 were adopted in our evaluation... 4http://perception.i2r.a-star.edu.sg/bkmodel/bkindex
Dataset Splits No The paper describes data generation and corruption (e.g., missing data, noise addition) and reports performance averaged over multiple runs or realizations, but it does not specify explicit training/validation/test dataset splits with percentages or counts for reproducibility.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., GPU/CPU models, memory) used for running experiments.
Software Dependencies No The paper refers to modifying existing solvers, but does not provide specific software names with version numbers for any libraries or dependencies used in their implementation.
Experiment Setup Yes Input: Incomplete data matrix Y Rm n with observation indexed by Ω, k0, kend, µ > 1. where parameter γ > 0 is introduced to control the strength of the weights assigned to the selected samples. the rank was set to 4 and 6 for rigid and nonrigid SFM, respectively.