Saved in:
Bibliographic Details
Main Authors: He, Zihong, Liang, Hai-Ning, Liang, Chen
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.14449
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917344874332160
author He, Zihong
Liang, Hai-Ning
Liang, Chen
author_facet He, Zihong
Liang, Hai-Ning
Liang, Chen
contents Response timing judgment is a critical component of interactive speech agents. Although there exists substantial prior work on turn modeling and voice wake-up, there is a lack of research on response timing judgments continuously aligned with user intent. To address this, we propose the Tap-to-Adapt framework, which enables users to naturally activate or interrupt the agent via tap interactions to construct online learning labels for response timing models. Under this framework, Dilated TCN and a sequential replay strategy play significant roles, as demonstrated through data-driven experiments and user studies. Additionally, we develop an evaluation and continuous data mining system tailored for the Tap-to-Adapt framework, through which we have collected approximately 20,000 samples from the user studies involving 20 participants.
format Preprint
id arxiv_https___arxiv_org_abs_2603_14449
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Tap-to-Adapt: Learning User-Aligned Response Timing for Speech Agents
He, Zihong
Liang, Hai-Ning
Liang, Chen
Human-Computer Interaction
Response timing judgment is a critical component of interactive speech agents. Although there exists substantial prior work on turn modeling and voice wake-up, there is a lack of research on response timing judgments continuously aligned with user intent. To address this, we propose the Tap-to-Adapt framework, which enables users to naturally activate or interrupt the agent via tap interactions to construct online learning labels for response timing models. Under this framework, Dilated TCN and a sequential replay strategy play significant roles, as demonstrated through data-driven experiments and user studies. Additionally, we develop an evaluation and continuous data mining system tailored for the Tap-to-Adapt framework, through which we have collected approximately 20,000 samples from the user studies involving 20 participants.
title Tap-to-Adapt: Learning User-Aligned Response Timing for Speech Agents
topic Human-Computer Interaction
url https://arxiv.org/abs/2603.14449