Generalized Conditional Gradient for Sparse Estimation
Authors: Yaoliang Yu, Xinhua Zhang, Dale Schuurmans
JMLR 2017 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | A series of experiments on matrix completion, multi-class classification, and multi-view dictionary learning shows that the proposed method can significantly reduce the training cost of current alternatives. |
| Researcher Affiliation | Academia | Yaoliang Yu EMAIL School of Computer Science University of Waterloo Waterloo, ON, N2L 3G1, Canada Xinhua Zhang EMAIL Department of Computer Science University of Illinois at Chicago Chicago, IL 60607 Dale Schuurmans EMAIL Department of Computing Science University of Alberta Edmonton, Alberta T6G 2E8, Canada |
| Pseudocode | Yes | Algorithm 1 Generalized Conditional Gradient (GCG). Algorithm 2 GCG for positively homogeneous regularizers. Algorithm 3 GCG variant for low rank learning. |
| Open Source Code | Yes | Open source Matlab code released at https://www.cs.uic.edu/~zhangx/GCG. |
| Open Datasets | Yes | Five data sets are used and their statistics are given in Table 1. The data sets Yahoo252M (Dror et al., 2012) and Yahoo60M8 provide music ratings, and the data sets Netflix (Dror et al., 2012), Movie Lens1M, and Movie Lens10M9 all provide movie ratings... We conducted experiments on three data sets extracted from the Image Net repository (Deng et al., 2009), with characteristics shown in Table 2 (Akata et al., 2014)... The data set we considered is based on the Extended Yale Face Database B13 (Georghiades et al., 2001). |
| Dataset Splits | Yes | For Movie Lens1M and Yahoo60M, we randomly sampled 75% ratings as training data, using the rest for testing. The other three data sets came with training and testing partition. |
| Hardware Specification | Yes | All experiments were run on a single core of a cluster housed at National ICT Australia with AMD 6-Core Opteron 4184 (2.8GHz, 3M L2/6M L3 Cache, 95W TDP, 64 GB memory). |
| Software Dependencies | No | All algorithms were implemented in Matlab unless otherwise noted... The local optimization in GCG was performed by L-BFGS with the maximum number of iteration set to 20.11 Downloaded from http://www.cs.ubc.ca/~pcarbo/lbfgsb-for-matlab.html. The paper mentions MATLAB and L-BFGS but does not provide specific version numbers for either, which is required for a reproducible description of ancillary software. |
| Experiment Setup | Yes | The regularization constant λ, given in Table 1, was chosen from {1, 10, 50, 100, 200, 300, 400, 500, 1000} to minimize the test RMSE... Here the value of λ was chosen from {10 4, 10 3, 10 2, 10 1} to maximize test accuracy... The local optimization in GCG was performed by L-BFGS with the maximum number of iteration set to 20. |