Online Kernel Selection via Incremental Sketched Kernel Alignment

Authors: Xiao Zhang, Shizhong Liao

IJCAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Empirical studies demonstrate that the proposed online kernel selection approach is computationally efficient while maintaining comparable accuracy for online kernel learning.
Researcher Affiliation Academia Xiao Zhang and Shizhong Liao School of Computer Science and Technology, Tianjin University, Tianjin 300350, China {xiaozhang, szliao}@tju.edu.cn
Pseudocode Yes Algorithm 1: OKS-ISKA Algorithm
Open Source Code No The paper does not provide concrete access to source code for the methodology described.
Open Datasets Yes We merged the training and testing data into a single dataset for each benchmark dataset4, and compared the proposed OKS-ISKA with the following state-of-the-art online kernel selection algorithms. 4http://www.csie.ntu.edu.cn/ cjlin/libsvmtools/datasets/ http://kdd.ics.uci.edu/databases/kddcup99/kddcup99.html
Dataset Splits No The paper mentions merging training and testing data and performing '20 different random permutations of the datasets' but does not provide specific percentages or counts for distinct training, validation, and test splits within these permutations for reproduction.
Hardware Specification Yes Algorithms were implemented in R 3.3.2 on a machine with 4-core Intel Core i7 3.60 GHz CPU and 16GB memory.
Software Dependencies Yes Algorithms were implemented in R 3.3.2 on a machine with 4-core Intel Core i7 3.60 GHz CPU and 16GB memory.
Experiment Setup Yes A set of Gaussian kernels with kernel widths σ {2 (i+1)/2, i = [ 12 : +2 : 12]} was adopted as the candidate kernel set and the kernel widths of OKL-GD were restricted to the same range. The initial parameter i of the kernel width was selected in { 12, 10, 8} uniformly... We tuned the stepsize of OGD in a range 10[ 5:+1:0] and the regularization parameter λ in a range 10[ 4:+1:1]. For our OKS-ISKA, we set q = 5 for the median estimate, µξ = 0.3, B = 150 for small datasets (T < 10, 000) and B = 200 for other datasets... Besides, we set ηt = 1/(tλ) as in Theorem 3.