Partial Label Learning with Dissimilarity Propagation guided Candidate Label Shrinkage
Authors: Yuheng Jia, Fuchao Yang, Yongqiang Dong
NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on artificial and real-world partial label data sets demonstrate the effectiveness of the proposed PLL method. |
| Researcher Affiliation | Academia | Yuheng Jia1, 3 , Fuchao Yang2, Yongqiang Dong1 1 School of Computer Science and Engineering, Southeast University, Nanjing 210096, China 2 College of Software Engineering, Southeast University Nanjing 210096, China 3 Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications (Southeast University), Ministry of Education, China yhjia@seu.edu.cn, yangfc@seu.edu.cn, dongyq@seu.edu.cn |
| Pseudocode | Yes | Algorithm 1 The Pseudo Code of the Proposed Method |
| Open Source Code | Yes | The code is publicly available at https://github.com/Yangfc-ML/DPCLS. |
| Open Datasets | Yes | To demonstrate the effectiveness of the proposed model, we compared DPCLS with eight shallow PLL algorithms, which were configured by the suggested parameters in the literature, i.e., CLPL [1], PL-SVM [12], PL-KNN [5], PL-DA [18], IPAL [22], AGGD [16], PL-CLA [13], SDIM [2]. Those methods were evaluated on 10 synthetic data sets and 7 real-world data sets, whose details can be found in Section C of the supplementary file. |
| Dataset Splits | Yes | Ten runs of 50%/50% random train/test splits were performed on each data set, and the average classification accuracy and the standard deviation were recorded. |
| Hardware Specification | No | The paper does not provide specific hardware details such as exact GPU/CPU models, processor types, or memory amounts used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with version numbers like Python 3.8, CPLEX 12.4) needed to replicate the experiment. |
| Experiment Setup | Yes | Hyper-parameter Settings of Our Method Parameter λ is used to control the model complexity. ... we set λ=0.05 for our method. Parameters α and β are used to control the importance of the adversarial term and dissimilarity propagation term respectively. According to a number of experiments, we fix β = 0.001 and select α from {0.001, 0.01}. Parameter k controls the number of k-nearest neighbors. Following the previous works [16, 22], we set k=10. |