Partial Label Learning with Self-Guided Retraining

Authors: Lei Feng, Bo An3542-3549

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on synthesized and real-world datasets demonstrate that the proposed approach significantly outperforms the state-of-the-art partial label learning approaches.
Researcher Affiliation Collaboration 1School of Computer Science and Engineering, Nanyang Technological University, Singapore 2Alibaba-NTU Singapore Joint Research Institute, Singapore
Pseudocode Yes The pseudo code of SURE is presented in Algorithm 1.
Open Source Code No The paper does not provide an explicit statement about the release of source code for the methodology described, nor does it include a link to a code repository.
Open Datasets Yes Following the widely-used controlling protocol (Cour, Sapp, and Taskar 2011; Liu and Dietterich 2012; Zhang and Yu 2015; Wu and Zhang 2018; Feng and An 2018; Wang and Zhang 2018), each UCI dataset can be used to generate artificial partial label datasets. These datasets are publicly available at: http://cse.seu.edu.cn/Personal Page/zhangml/
Dataset Splits Yes Parameters for each algorithm are selected by five-fold cross-validation on the training set. For each dataset, ten-fold cross-validation is performed where mean prediction accuracies and the standard deviations are recorded.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU/GPU models, memory, or cloud instance types) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details, such as library names with version numbers, needed to replicate the experiment.
Experiment Setup Yes The two parameters λ and β for SURE are chosen from {0.001, 0.01, 0.05, 0.1, 0.3, 0.5, 1}. Parameters for each algorithm are selected by five-fold cross-validation on the training set. In this paper, Gaussian kernel function κ(xi, xj) = exp( xi xj 2 2 /(2σ2)) is employed with σ set to the averaged pairwise distances of instances.