Low-Complexity DNN-Based End-to-End Automatic Speech Recognition Using Low-Rank Approximation
International SoC Design Conference(2020)
摘要
Targeting the on-device speech-to-text application for streaming inputs, this paper presents an efficient way to reduce the computational complexity of deep neural networks (DNNs) for attention-based speech processing. The proposed technique applies the singular value decomposition (SVD) to the large-sized matrix multiplications, removing less important computations by utilizing the low-rank approximation. The clipping thresholds are carefully adjusted to relax the computing costs as well as the memory overheads while maintaining the recognition accuracy.
更多查看译文
关键词
Speech recognition,on-device DNN processing,low-complexity AI algorithm,low-rank approximation
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要