Multi-Dimensional Classification via Sparse Label Encoding

Authors: Bin-Bin Jia, Min-Ling Zhang

ICML 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results clearly validate the superiority of SLEM against state-of-the-art MDC approaches. In this paper, the experiments are conducted over a total of 11 benchmark data sets, whose detailed characteristics are summarized in Table 1, including the number of examples (#Exam.), the number of dimensions (#Dim.), the number of class labels per dimension (#Labels/Dim.),4 and the number of features (#Features).
Researcher Affiliation Academia 1School of Computer Science and Engineering, Southeast University, Nanjing 210096, China 2College of Electrical and Information Engineering, Lanzhou University of Technology, Lanzhou 730050, China 3Key Lab. of Computer Network and Information Integration (Southeast University), Ministry of Education, China.
Pseudocode Yes Algorithm 1 Solving problem (6) via APG. Algorithm 2 LOMP: v = R(z, A, k, I)
Open Source Code No The paper does not provide any statement or link indicating that the source code for the methodology is openly available.
Open Datasets Yes In this paper, the experiments are conducted over a total of 11 benchmark data sets, whose detailed characteristics are summarized in Table 1, including the number of examples (#Exam.), the number of dimensions (#Dim.), the number of class labels per dimension (#Labels/Dim.),4 and the number of features (#Features).
Dataset Splits Yes In the experiments, we conduct ten-fold cross validation over each data set for all compared approaches, and both mean metric value and standard deviation are recorded for performance comparison.
Hardware Specification No The paper mentions 'We thank the Big Data Center of Southeast University for providing the facility support on the numerical calculations in this paper' but does not provide specific hardware details like GPU/CPU models or configurations.
Software Dependencies No Specifically, the Libsvm package (Chang & Lin, 2011) with default parameter settings is used in experiments. The paper mentions a software package but does not provide specific version numbers for it or other software dependencies.
Experiment Setup Yes For g MML, its parameters λ, t, γ and k are tuned as suggested in (Ma & Chen, 2018). For the proposed SLEM approach, we use random Gaussian matrix to serve as the encoding matrix A with s = s 1, and the three trade-off parameters in the formulation (2) are set as λ = 1, γ1 = 1 and γ2 = 1.