Theory of Dual-sparse Regularized Randomized Reduction
Authors: Tianbao Yang, Lijun Zhang, Rong Jin, Shenghuo Zhu
ICML 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In numerical experiments, we present an empirical study on a real data set to support our analysis and we also demonstrate a novel application of the reduction and recovery framework in distributed learning from LSHD data. and 5. Numerical Experiments In this section, we provide a case study in support of DSRR and the theoretical analysis, and a demonstration of the application of DSRR to distributed optimization. |
| Researcher Affiliation | Collaboration | Tianbao Yang TIANBAO-YANG@UIOWA.EDU Department of Computer Science, the University of Iowa, Iowa City, USA Lijun Zhang ZHANGLJ@@LAMDA.NJU.EDU.CN National Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China Rong Jin RONGJIN@CSE.MSU.EDU Department of Computer Science and Engineering, Michigan State University, East Lansing, USA Institute of Data Science and Technologies at Alibaba Group, Seattle, USA Shenghuo Zhu SHENGHUO@GMAIL.COM Institute of Data Science and Technologies at Alibaba Group, Seattle, USA |
| Pseudocode | No | The paper describes methods and procedures in narrative text and mathematical formulations but does not include structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any statement or link regarding the open-sourcing of code for the described methodology. |
| Open Datasets | Yes | We use the RCV1binary data (Lewis et al., 2004) to conduct a case study. The data contains 697, 641 documents and 47, 236 features. We use a splitting 677, 399/20, 242 for training and testing. and Two data sets are used, namely RCV1-binary, KDD 2010 Cup data. For KDD 2010 Cup data, we use the one available on Lib SVM data website. |
| Dataset Splits | No | The paper mentions a training/testing split for RCV1 data (“677, 399/20, 242 for training and testing”) but does not specify a validation set or explicit splits for the KDD data. |
| Hardware Specification | No | The paper does not provide specific details about the hardware used for running experiments, such as CPU or GPU models. |
| Software Dependencies | No | The paper mentions software like Vowpal Wabbit and the distributed stochastic dual coordinate ascent (Dis DCA) algorithm, but it does not specify any version numbers for these or other software dependencies. |
| Experiment Setup | Yes | We vary the value of τ among 0, 0.1, 0.2, . . . , 0.9, the value of m among 1024, 2048, 4096, 8192, and the value of λ among 0.001, 0.00001. and The value of λ = 10 5 and the value of τ = 0.9. The high-dimensional features are reduced to m = 1024-dimensional space using random hashing. The loss function is the squared hinge loss. |