Universal Matrix Completion
Authors: Srinadh Bhojanapalli, Prateek Jain
ICML 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we will present a few empirical results on both synthetic and real data sets. The goal of this section is to demonstrate effect of the spectral gap of the sampling graph G (associated with ) on successful recovery of a matrix. First, we use synthetic data sets generated in the following manner. We then generate rank-10 matrix M, using M = UV T. As U, V are sampled from the normal distribution, hence w.h.p., M satisfies incoherence assumptions A1, A2 mentioned in Section 3. Next, we generate a sequence of sampling operators P (and the associated graph G) with varying (relative) spectral gap(1 σ2(G)/σ1(G)) by using a stochastic block model. |
| Researcher Affiliation | Collaboration | Srinadh Bhojanapalli BSRINADH@UTEXAS.EDU The University of Texas at Austin Prateek Jain PRAJAIN@MICROSOFT.COM Microsoft Research, India |
| Pseudocode | No | The paper presents mathematical formulations of problems and theoretical proofs but does not include 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 described is open-source or publicly available. |
| Open Datasets | Yes | Finally we take a real dataset of temperature values(T) for 365 days at 316 different locations from (NCDC), which has been used to test matrix completion algorithms (Candes & Plan, 2010) before. NCDC. National climatic data center. URL http:// www.ncdc.noaa.gov/oa/ncdc.html. |
| Dataset Splits | No | The paper describes generating synthetic data and sampling entries from a real dataset but does not specify any training, validation, or test splits for reproduction. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory) used to run the experiments. |
| Software Dependencies | No | The paper mentions using an 'Augmented Lagrangian Method (ALM) based method (Lin et al., 2010)' but does not specify any software libraries or their version numbers used in the implementation. |
| Experiment Setup | No | The paper describes how synthetic data and sampling operators are generated, and defines success criteria, but it does not provide specific hyperparameter values or optimizer settings for the Augmented Lagrangian Method (ALM) used to solve the nuclear norm minimization problem. |