Learning From Semi-Supervised Weak-Label Data
Authors: Hao-Chen Dong, Yu-Feng Li, Zhi-Hua Zhou
AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments validate the effectiveness of SSWL. In this section, we first give the experimental setup and then show the evaluation of our proposal compared to several state-of-the-art algorithms on a number of real-world tasks. |
| Researcher Affiliation | Academia | Hao-Chen Dong, Yu-Feng Li, Zhi-Hua Zhou National Key Laboratory for Novel Software Technology, Nanjing University, Nanjing 210023, China Collaborative Innovation Center of Novel Software Technology and Industrialization, Nanjing, 210023, China {donghc, liyf, zhouzh}@lamda.nju.edu.cn |
| Pseudocode | Yes | Algorithm 1: SSWL Method |
| Open Source Code | No | The paper does not provide any statement or link indicating that the source code for the SSWL method is openly available. |
| Open Datasets | Yes | The text classification task is collected from SIAM Text Mining Competition (TMC). TMC dataset (Srivastava and Zane-Ulman 2005) is a large text dataset... The Yeast data set (Elisseeff and Weston 2001)is a gene function classification dataset... The Scene Image data set (Zhang and Zhou 2007) contains 2,000 natural scene images... We use the Microsoft Research image annotation data set (msrc) to verify that our method can help predict. |
| Dataset Splits | Yes | We randomly selected 1500 instances for training (500 labeled and 1000 unlabeled) and used the rest for testing. For all the methods, we conducted parameter selection for each evaluation metric by performing 5-fold crossvalidation on the training set. |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., GPU models, CPU types, memory) used to run the experiments. |
| Software Dependencies | No | LIBSVM (Chang and Lin 2011) package is employed for the implementation for the BSVM method and RBF kernel with the recommended parameter is employed. However, no specific version number for LIBSVM or other software dependencies is provided. |
| Experiment Setup | Yes | For our SSWL method and the SSWL-wo method, 5 nearest neighbor graph is used for the instance matrix in all the experiments. For our approach, we selected the trade-off parameters α, β and ζ from {10−2, . . . , 102}. |