Learning Discriminative Correlation Subspace for Heterogeneous Domain Adaptation
Authors: Yuguang Yan, Wen Li, Michael Ng, Mingkui Tan, Hanrui Wu, Huaqing Min, Qingyao Wu
IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Comprehensive experiments on two real-world data sets demonstrate the effectiveness of the proposed method compared to the state-of-the-art methods. |
| Researcher Affiliation | Academia | Yuguang Yan1 , Wen Li2 , Michael Ng3, Mingkui Tan1, Hanrui Wu1, Huaqing Min1, and Qingyao Wu1 1School of Software Engineering, South China University of Technology, China 2Computer Vision Laboratory, ETH Zurich, Switzerland 3Department of Mathematics, Hong Kong Baptist University, Hong Kong, China yan.yuguang@mail.scut.edu.cn, liwen@vision.ee.ethz.ch, qyw@scut.edu.cn |
| Pseudocode | Yes | Algorithm 1 Discriminative Canonical Correlation Analysis |
| Open Source Code | No | The paper mentions a third-party tool 'minfunc' with a link, but does not provide concrete access to the authors' own source code for the methodology described in the paper. |
| Open Datasets | Yes | Office Data Set. The Office data set [Saenko et al., 2010]...Office-Caltech Data Set. The Caltech-256 data set [Griffin et al., 2007] |
| Dataset Splits | No | The paper mentions using labeled target instances for training and testing ('randomly choose 3 target instances per category as the labeled target data for all the target domains, and the rest instances as the test data') but does not explicitly define a separate validation split or dataset. |
| Hardware Specification | No | The paper does not provide specific details about the hardware used to run the experiments, such as CPU/GPU models or memory specifications. |
| Software Dependencies | No | The paper mentions using the 'LBFGS algorithm' and refers to 'minfunc: unconstrained differentiable multivariate optimization in matlab' (Schmidt, 2005), but does not provide specific version numbers for MATLAB or any other software libraries required for reproducibility. |
| Experiment Setup | Yes | For our DCA method, we empirically set the subspace dimension as d C = 20, the trade-off parameter as C = 1, and the penalty parameter of ADMM as ρ = 5. |