A Batch-to-Online Transformation under Random-Order Model

Authors: Jing Dong, Yuichi Yoshida

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 9 Experiments We here provide a preliminary empirical evaluation of our framework in the context of online k-means clustering, and online linear regression, with the result shown in Figure 1. Our experiments are conducted with various approximation ratios and experimental setups (ϵ = 0.1, 0.01, 0.001, with k = 3 or k = 5 clusters). We then compare the performance of the proposed algorithm to the hindsight optimal solution. For k-means clustering, we obtain the hindsight optimal solution by applying k-means++ to all the data. In the context of regression, we utilize the least square formula to compute the hindsight optimal solution. Our experimental results demonstrate that the proposed algorithm is highly effective, and its performance aligns with our theoretical findings.
Researcher Affiliation Academia Jing Dong The Chinese University of Hong Kong, Shenzhen jingdong@link.cuhk.edu.cn Yuichi Yoshida National Institute of Informatics yyoshida@nii.ac.jp
Pseudocode Yes Algorithm 1: General batch-to-online conversion
Open Source Code No Not found. The paper does not provide any explicit statements or links about providing open-source code for their methodology.
Open Datasets No Not found. The paper mentions using 'online k-means clustering, and online linear regression' datasets in Section 9, but does not specify which datasets were used, nor does it provide any links, DOIs, citations, or explicit statements about their public availability. It just refers to the 'data' in general, for example, 'all the data' or 'aggregated data'.
Dataset Splits No Not found. The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) needed to reproduce the data partitioning.
Hardware Specification No Not found. The paper does not provide any specific hardware details such as GPU/CPU models or types of machines used for experiments.
Software Dependencies No Not found. The paper does not provide any specific software versions or dependencies.
Experiment Setup Yes Our experiments are conducted with various approximation ratios and experimental setups (ϵ = 0.1, 0.01, 0.001, with k = 3 or k = 5 clusters).