Self-Paced Active Learning: Query the Right Thing at the Right Time

Authors: Ying-Peng Tang, Sheng-Jun Huang5117-5124

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results show that the proposed approach is superior to state-of-the-art batch mode active learning methods.
Researcher Affiliation Academia College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics Collaborative Innovation Center of Novel Software Technology and Industrialization Nanjing 211106, China {tangyp, huangsj}@nuaa.edu.cn
Pseudocode Yes Algorithm 1 The SPAL Algorithm
Open Source Code No The paper mentions using external tools like CVX and MOSEK, providing their URLs: 'CVX (Grant and Boyd 2014) and MOSEK 1 are used to solve the QP problem. 1http://www.mosek.com/ http://cvxr.com/cvx'. However, there is no explicit statement about releasing the source code for their own proposed method (SPAL) or a link to its repository.
Open Datasets Yes We perform the experiments on 9 datasets, whose sizes are summarized in Table 1. Dataset thyroid antivirus clean1 # Instances 215 373 476 Dataset tictactoe image krvskp # Instances 958 2086 3196 Dataset phoneme gisette phishing # Instances 5404 7000 11055
Dataset Splits Yes For each dataset, we randomly sample 40% instances as the test set, and the rest 60% instances for the training. Further, 5% of the training set is used as the initially labeled data, while the rest instances consist of the unlabeled pool for active selection.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory specifications, or cloud computing instance types) used for running the experiments.
Software Dependencies No The paper mentions 'CVX (Grant and Boyd 2014) and MOSEK 1 are used to solve the QP problem' but does not specify their version numbers or other software dependencies with specific versions.
Experiment Setup Yes For the proposed method SPAL, we fix µ = 0.1, and γ = 0.1. For the SPL parameter λ, we initialize it with a certain value which is selected from {0.1, 0.01}, and follow the method used in (Lin et al. 2018) to update it linearly with a small fixed value. In our experiments, we fix λpace = 0.01 for all datasets. Specifically, we have the following updating rule for λ at tth iteration: λt = λinitial + (t - 1) λpace. We fix batch size b = 5 for all methods. The parameters of BMDR are set to the recommended values in their paper. Specifically, the regularized weight γ = 0.1 and the trade-off parameter µ = 1000.