http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
조준서(Junseo Jo),이영주(Youngjoo Lee) 대한전자공학회 2020 대한전자공학회 학술대회 Vol.2020 No.8
This paper presents several pruning methods to optimize the capsule network, which generally requires an excessive amount of computational costs compared to the conventional convolutional neural network. In order to realize the cost—effective but accurate solution, we newly introduce an advanced pruning technique dedicated to the capsule network architecture, handling only the length of capsule vectors while preserving its direction. Experimental results show that the proposed method allows to prune the given capsule network aggressively while maintaining the recognition accuracy compared to the naïve pruning approach.