10,000+ Times Accelerated Robust Subset Selection

Authors: Feiyun Zhu, Bin Fan, Xinliang Zhu, Ying Wang, Shiming Xiang, Chunhong Pan

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on ten benchmark datasets verify that our method not only outperforms state of the art methods, but also runs 10,000+ times faster than the most related method.
Researcher Affiliation Academia Institute of Automation, Chinese Academy of Sciences {fyzhu, bfan, ywang, smxiang and chpan}@nlpr.ia.ac.cn, zhuxinliang2012@ia.ac.cn
Pseudocode Yes Algorithm 1 for (13): A = ARSSA (X, V, P, IL, β) Input: X, V, P, IL, β 1: if N L then 2: update A via the updating rule (14), that is 3: A = β V + βXT X 1XT P. 4: else if N > L then 5: update A via the updating rule (15), that is 6: A = B (IL + XB) 1P, where B = β XV 1 T . 7: end if Output: A
Open Source Code No The paper does not provide any concrete access to source code for the described methodology.
Open Datasets Yes Brief descriptions of ten benchmark datasets are summarized in Table 2, where Total(N ) denotes the total set of samples in each data.
Dataset Splits No The paper states: 'The top 200 representative samples are selected for training.' and 'The remainder (except candidate set) are used for test.' but does not explicitly provide details about validation splits, percentages, or the methodology for partitioning data into training, validation, and test sets.
Hardware Specification Yes All experiments are conducted on a server with 64-core Intel Xeon E7-4820 @ 2.00 GHz, 18 Mb Cache and 0.986 TB RAM, using Matlab 2012.
Software Dependencies Yes using Matlab 2012.
Experiment Setup No The paper mentions general experimental settings like selecting the top 200 representative samples for training, but it does not provide specific hyperparameter values (e.g., learning rate, batch size) or detailed system-level training configurations.