Signal and Noise Statistics Oblivious Orthogonal Matching Pursuit
Authors: Sreejith Kallummil, Sheetal Kalyani
ICML 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Both analytical results and numerical simulations in real and synthetic data sets indicate that RRT has a performance comparable to OMP with a priori knowledge of sparsity and noise statistics. ... Section 5 presents numerical simulation results. |
| Researcher Affiliation | Academia | 1Department of Electrical Engineering, IIT Madras, India 2Department of Electrical Engineering, IIT Madras, India. |
| Pseudocode | Yes | Algorithm 1 Orthogonal matching pursuit; Algorithm 2 Residual ratio thresholding |
| Open Source Code | No | The paper does not contain any statement about making its source code available, nor does it provide any links to a code repository. |
| Open Datasets | Yes | We consider four widely studied real life data sets and compare the outliers identified by these algorithms with the existing and widely replicated studies on these data sets. More details on these data sets are given in the supplementary materials. ... Data Set Outliers reported in literature RRT CV LAT Stack Loss ... (Rousseeuw & Leroy, 2005) ... AR2000 ... (Atkinson & Riani, 2012) ... Brain Body Weight ... (Rousseeuw & Leroy, 2005) ... Stars ... (Rousseeuw & Leroy, 2005) |
| Dataset Splits | No | The paper describes synthetic data generation and mentions real-life datasets, but it does not specify any training, validation, or test splits for these datasets. It refers to 'five fold CV' in the context of tuning a baseline OMP, not for the data partitioning of its own experiments. |
| Hardware Specification | No | The paper does not provide any specific details regarding the hardware (e.g., CPU, GPU models, memory) used to conduct the numerical simulations or experiments. |
| Software Dependencies | No | The paper discusses various algorithms and concepts but does not list any specific software libraries, frameworks, or their version numbers used for implementing the methods or running experiments. |
| Experiment Setup | Yes | The synthetic data sets are generated as follows. ... The noise w is sampled according to N(0n, σ2In) with σ2 = 1. The non zero entries of β are randomly assigned βj = 1. Subsequently, these entries are scaled to achieve SNR = Xβ 2 2/n = 3. The number of non zero entries k0 in all experiments are fixed at six. ... RRT1 and RRT2 represent RRT with parameter α set to α = 1/ log(n) and α = 1/ n respectively. ... we set kmax = min(p, [0.5(rank(X) + 1)]). |