Title :
Incremental new actions learning system with limited cost and storage
Author :
Che-Wei Chang ; Liang-Gee Chen
Author_Institution :
Grad. Inst. of Electron. Eng., Nat. Taiwan Univ., Taipei, Taiwan
Abstract :
In the future robotic applications, robot requires the ability not only to recognize human actions but also to learn incrementally and quickly. Therefore, we proposed an incremental action learning system for this future requirement. The proposed system can continuously learn new actions quickly with robust performance and less effort.
Keywords :
human-robot interaction; learning (artificial intelligence); robot programming; robust control; future robotic applications; human actions; incremental new actions learning system; robust performance; Accuracy; Consumer electronics; Learning systems; Robots; Robustness; Support vector machines; Visualization; Action recognition; Incremental learning; Nearest class mean classification; new class learning; representative data selection;
Conference_Titel :
Consumer Electronics (ISCE), 2015 IEEE International Symposium on
Conference_Location :
Madrid
DOI :
10.1109/ISCE.2015.7177847
