GALAXY: A Generative Pre-trained Model for Task-Oriented Dialog with Semi-Supervised Learning and Explicit Policy Injection.
Wanwei HeYinpei DaiYinhe ZhengYuchuan WuZheng CaoDermot LiuPeng JiangMin YangFei HuangLuo SiJian SunYongbin LiPublished in: CoRR (2021)