Nonparametric Estimation of Multi-View Latent Variable Models
Authors: Le Song, Animashree Anandkumar, Bo Dai, Bo Xie
ICML 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In both synthetic and real world datasets, the nonparametric tensor power method compares favorably to EM algorithm and other spectral algorithms. Experimentally, we corroborate our theoretical results by comparing our algorithm to the EM algorithm and previous spectral algorithms. |
| Researcher Affiliation | Academia | Le Song LSONG@CC.GATECH.EDU Georgia Institute of Technology, Atlanta, GA 30345 USA Animashree Anandkumar A.ANANDKUMAR@UCI.EDU University of California, Irvine, CA 92697, USA Bo Dai, Bo Xie BODAI,BXIE33@GATECH.EDU Georgia Institute of Technology, Atlanta, GA 30345 USA |
| Pseudocode | Yes | The overall kernel algorithm is summarized in Algorithm 1. The tensor power method is provided in the Appendix in Algorithm 2 for completeness. |
| Open Source Code | No | The paper does not contain any explicit statements or links indicating that the source code for the described methodology is publicly available. |
| Open Datasets | Yes | We use the DLBCL Lymphoma dataset collection from (Aghaeepour et al., 2013) to compare our kernel algorithm with the four alternatives. This collection contains 24 datasets with two or three clusters, and each dataset consists of tens of thousands of cell measurements in 5 dimensions. |
| Dataset Splits | Yes | For each dataset, we select the best kernel bandwidth by 5-fold cross validation using log-likelihood. |
| Hardware Specification | No | The paper does not provide specific details about the hardware used for running the experiments (e.g., GPU models, CPU types, or cloud instance specifications). |
| Software Dependencies | No | The paper discusses various algorithms (e.g., EM algorithm, k-means, spectral algorithms) but does not provide specific version numbers for any software libraries, programming languages, or solvers used in its implementation. |
| Experiment Setup | Yes | For each dataset, we select the best kernel bandwidth by 5-fold cross validation using log-likelihood. The mixture proportion for the h-th component is set to be h = 2h (k+1), 8h 2 [k] (unbalanced). The EM algorithm is not guaranteed to find the global solution in each trial. Thus we randomly initialize it 10 times. |