TY  - RPRT
TI  - D2: High-Capability Full-Duplex Spoken Language Models from Half-Duplex Backbones
AU  - Cheng-Kuang Chang
AU  - Alexander H. Liu
AU  - Brian Kingsbury
AU  - Hilde Kuehne
AU  - Rogerio Feris
AU  - Samuel Thomas
AU  - James Glass
AU  - Kai-Wei Chang
PY  - 2026
UR  - https://arxiv.org/abs/2610.17124
ID  - 2610.17124
ER  -