Multi-timescale Feature-extraction Architecture of Deep Neural Networks for Acoustic Model Training from Raw Speech Signal.

Ryu Takeda, Kazuhiro Nakadai, Kazunori Komatani

Browse the full IROS paper archive.