Compressive Sensing of Signals from a GMM with Sparse Precision Matrices

Authors: Jianbo Yang, Xuejun Liao, Minhua Chen, Lawrence Carin

NeurIPS 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The proposed method is demonstrated extensively on compressive sensing of imagery and video, and the results with simulated and hardware-acquired real measurements show significant performance improvement over state-of-the-art methods.
Researcher Affiliation Academia 1Department of Electrical and Computer Engineering, Duke University 2Department of Statistics & Department of Computer Science, University of Chicago
Pseudocode No The paper describes the steps of its variational EM algorithm but does not provide a formal pseudocode block or algorithm listing.
Open Source Code No The paper mentions a website for complete results and other settings, but it does not explicitly state that the source code for the methodology is available there or elsewhere.
Open Datasets Yes The USPS handwritten digits dataset 3 and the face dataset [28] are used in this experiment. It is downloaded from http://cs.nyu.edu/ roweis/data.html.
Dataset Splits No The paper describes the datasets used and the number of images/frames, but it does not provide specific training, validation, or test split percentages, counts, or explicit instructions for reproducibility.
Hardware Specification No The paper mentions 'hardware-acquired real measurements' and 'CACTI camera [6]' which refer to the data acquisition hardware, not the computing hardware (CPUs, GPUs, etc.) used for running the experiments.
Software Dependencies No The paper mentions various methods/algorithms (e.g., KSVD-OMP, Tw IST, GAP) and a 'single pixel camera' setup, but it does not list specific software dependencies with version numbers.
Experiment Setup Yes For the proposed methods, the hyperparameters of the scaled mixture of Gaussians are set as pa0/b0/N ≈ 300, c0 = d0 = 10^-6, the hyperparameter of Dirichlet prior α0 is set as a vector with all elements being one, the hyperparameters of the mean of each Gaussian component are set as β0 = 1, and m0 is set to the mean of the initialization of {bxi}N i=1. We fixed κ = 10^-6 for the proposed methods, GMM-TP and PLE.