Guaranteed Tensor Decomposition: A Moment Approach

Authors: Gongguo Tang, Parikshit Shah

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

Reproducibility Variable Result LLM Response
Research Type Experimental Numerical experiments are conducted to test the performance of the moment approach. We performed a series of experiments to illustrate the performance of the SDP relaxations (9) in solving the tensor decomposition and other related problems.
Researcher Affiliation Collaboration Gongguo Tang GTANG@MINES.EDU Colorado School of Mines, 1500 Illinois Street, Golden, CO, USA Parikshit Shah PSHAH@DISCOVERY.WISC.EDU Yahoo! Labs, 701 First Ave, Sunnyvale CA, USA
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any statement or link indicating that the source code for the methodology is openly available.
Open Datasets No The paper describes generating synthetic data for experiments (e.g., 'We produced T = 10 instances', 'A set of r random, orthonormal vectors... were generated to produce the tensor A') but does not specify or provide access information for any publicly available or open dataset.
Dataset Splits No The paper does not explicitly provide information about train/validation/test dataset splits. It discusses generating synthetic data and sampling observations for completion tasks, but not in terms of distinct training, validation, and testing sets.
Hardware Specification No The paper does not provide specific details regarding the hardware used for running the experiments. It does not mention CPU, GPU, or memory specifications.
Software Dependencies No The paper mentions 'All the SDPs are solved using the CVX package.' but does not specify a version number for CVX or any other software dependencies.
Experiment Setup Yes In preparing the upper plot in Figure 1, we took n = 10, r 2 {2, 4, . . . , 20}, and 2 {0.38, 0.39, . . . , 0.52}. Gaussian noise of standard deviation σ equal to half the average magnitude of the tensor elements was added to all the unique entries of the tensor. We then ran the optimization (22) with k = 2 to perform denoising. The penalization parameter γ is set to equal σ.