Reproduce “AdaptDuplex: From Static to Adaptive Full-Duplex Spoken Dialogue” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhiyang Zhou, Yingxin Shang, Zhou Wang, Hongwei Cai, Weixu Wang, Shuran Zhou, Shuofeng Zhao, Wenke Fan and 4 more
AdaptDuplex enables full-duplex speech by letting a model choose among discrete listening-window durations and take explicit interaction actions while speaking. Its Thinker–Talker protocol supports asynchronous memory or external reasoning, and tests compare dynamic with fixed windows. On a controlled test, dynamic windows nearly match the long-window backchannel score with lower latency, while short windows retain better interruption behavior and latency.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.