TY  - RPRT
TI  - A Human-Inspired Decoupled Architecture for Efficient Audio Representation Learning
AU  - Harunori Kawano
AU  - Takeshi Sasaki
PY  - 2026
UR  - https://arxiv.org/abs/2603.26098
ID  - 2603.26098
ER  -