DuplexGen: Adaptive Synthesis of Human–AI Turn-Taking Dialogues

Takyoung Kim^1*, Kang-wook Kim^2,4*, Sang Hoon Woo^2,5, Julia Hirschberg³, Gunhee Kim², Dilek Hakkani-Tür¹

¹University of Illinois Urbana-Champaign   ²Seoul National University   ³Columbia University   ⁴University of California, Berkeley   ⁵Georgia Institute of Technology

^*Equal contribution

📄 Paper  ·  🌐 Samples & Demo  ·  💻 Code  ·  🤗 Corpus  ·  🤗 Spoken

---

Code release coming soon. The full code will be released by the end of the first week of August.

This repository will host the recipe for fine-tuning a full-duplex model (PersonaPlex) on DuplexGen-generated data.

---

Abstract

Turn-taking is a central component of full-duplex interaction. Which turn-taking behaviors are appropriate varies with the scenario, yet current models apply a single norm regardless of context. This limitation originates in their training data: human–human speech corpora capture natural timing phenomena but provide little role grounding or scenario-specific norms, while heuristic or prompted synthesis methods inject turn-taking behaviors without basing them on human preferences. We introduce DuplexGen, a framework for generating dialogues with scenario-adaptive turn-taking by calibrating LLM predictions against a small set of slot-level human preference annotations. In six cooperative and competitive tasks, human turn-taking preferences differ systematically, and DuplexGen aligns substantially more closely with those preferences than uncalibrated prompting or training solely on generic human–human data; a full-duplex model trained on DuplexGen-generated data exhibits distinctive, human-preferred turn-taking behaviors. These results show that human calibration, not corpus scale or prompt design alone, is what allows turn-taking synthesis to be scenario-specific.

Citation

@misc{kim2026duplexgenadaptivesynthesishumanai,
      title={DuplexGen: Adaptive Synthesis of Human-AI Turn-Taking Dialogues},
      author={Takyoung Kim and Kang-wook Kim and Sang Hoon Woo and Julia Hirschberg and Gunhee Kim and Dilek Hakkani-Tür},
      year={2026},
      eprint={2607.26178},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2607.26178},
}