Title :
I don´t know the label: Active learning with blind knowledge
Author :
Meng Fang ; Xingquan Zhu
Author_Institution :
Centre for Quantum Comput. & Intell. Syst., Univ. of Technol. Sydney, Sydney, NSW, Australia
Abstract :
Active learning traditionally assumes that the oracle is capable of providing labeling information for each query instance. In reality, the oracle might have no information for some queries and cannot provide accurate label but only answers “I don´t know the label”. We focus on this problem and provide a unified objective function to ensure that each query instance submitted to the oracle is the one mostly needed for labeling and the oracle should also have sufficient knowledge to label. Experimental results on real-world and benchmark data sets demonstrate the effectiveness of the proposed design for supporting active learning using oracles with blind knowledge.
Keywords :
image classification; learning (artificial intelligence); query processing; active learning; benchmark data sets; blind knowledge; labeling information; oracle; query instance; real-world data sets; unified objective function; Accuracy; Benchmark testing; Entropy; Knowledge based systems; Labeling; Linear programming; Uncertainty;
Conference_Titel :
Pattern Recognition (ICPR), 2012 21st International Conference on
Conference_Location :
Tsukuba
Print_ISBN :
978-1-4673-2216-4