Beyond Unfolding: Exact Recovery of Latent Convex Tensor Decomposition Under Reshuffling

Authors: Chao Li, Mohammad Emtiyaz Khan, Zhun Sun, Gang Niu, Bo Han, Shengli Xie, Qibin Zhao4602-4609

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on synthetic data validate our theory, and results on image steganography show that our method outperforms the state-of-the-art methods.
Researcher Affiliation Academia Chao Li,1 Mohammad Emtiyaz Khan,1 Zhun Sun,1 Gang Niu,1 Bo Han,1 Shengli Xie,2, Qibin Zhao1, 1RIKEN Center for Advanced Intelligence Project (AIP), Japan 2School of Automation, Guangdong University of Technology, China
Pseudocode Yes Algorithm 1 Reshuffled-TD
Open Source Code Yes Supplementary materials are available at: http://qibinzhao.github.io.
Open Datasets Yes The datasets we used in the experiment include texture (DTD), natural (LIVE and FIVEK (Bychkovsky et al. 2011)), cartoon (CART. (Royer et al. 2017)) and fingerprint (FVC (Maltoni et al. 2009)) datasets.
Dataset Splits No The paper does not explicitly provide training/validation/test dataset splits with specific percentages or sample counts.
Hardware Specification No No specific hardware details (e.g., CPU, GPU models, memory) used for running experiments are mentioned in the paper.
Software Dependencies No No specific software dependencies or their version numbers (e.g., Python, PyTorch, specific libraries) are mentioned in the paper.
Experiment Setup Yes We generate data by using N square matrices A i Rn n, i [N]. Each A i is generated by multiplying two random semiorthonormal matrices with rank r... We fix the size of the components n = 100, the number of the components N = 10 and set the rank of each component by r = 1, . . . , 4. Then, we add the zero-mean i.i.d. Gaussian noise to the data, and the variance of the noise is controlled by the signal to noise ratio (SNR). During the concealing phase, we consider each channel of the secret image as one component, and they are randomly reshuffled. To estimate the number of components by Reshuffled-TD, we compare the norm of the recovered components with a threshold (we choose η = 0.1 for numerical consideration).