Reproduce “DuplexGen: Adaptive Synthesis of Human-AI Turn-Taking Dialogues” 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
Takyoung Kim, Kang-wook Kim, Sang Hoon Woo, Julia Hirschberg, Gunhee Kim, Dilek Hakkani-Tür
DuplexGen trains assistant turn-taking for six cooperative and competitive tasks by calibrating action predictions at partial-utterance slots against distributions of human preferences, then using the predictor to synthesize spoken-style dialogue. For Qwen3-4B, its combined Switchboard and calibration method reaches a six-task average KL divergence of 0.335, below prompt-only and Switchboard-only; results cover evaluated scenarios, not all roles.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.