@inproceedings{li-etal-2026-dual,
title = "Dual-Reasoner: Bridging Interleaved Atomicity and Streaming Latency via Thinking-while-Talking",
author = "Li, Yangzhuo and
Ji, Shengpeng and
Chen, Yifu and
Liang, Tianle and
Yang, Haoyu and
Junboli and
Fang, Jun and
Li, Lin and
Hong, Qingyang",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.199/",
pages = "4081--4105",
ISBN = "979-8-89176-395-1",
abstract = "Integrating explicit Chain-of-Thought (CoT) into end-to-end spoken dialogue models enhances intelligence but incurs prohibitive latency. While the ``Thinking-while-Talking'' paradigm alleviates this delay, it fundamentally compromises block atomicity, severing the logical connection between interleaved thought and speech. To address this, we present Dual-Reasoner, employing a Streaming Masking Mechanism underpinned by our Dual-Think-30k dataset to guarantee uninterrupted audio streaming. Crucially, to strictly align the fragmented thinking blocks to service speech generation, we introduce the Atomic-Consistency Restoration framework. To secure comprehensive capabilities in high-difficulty reasoning, this mechanism utilizes a quadruple-constraint system to reconstruct logical atomicity, ensuring that ``think'' chunks act as a rigorous anchor for ``talk'' outputs. Experimental results demonstrate that Dual-Reasoner achieves comprehensive reasoning enhancements within ultra-low latency constraints: it elevates the VoiceBench score from 67.24 to 73.41 over the baseline, while significantly reducing the Time-to-First-Audio (TTFA) from 20.35s to 3.65s and the Real-Time Factor (RTF) from 7.04 to 1.05."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2026-dual">
<titleInfo>
<title>Dual-Reasoner: Bridging Interleaved Atomicity and Streaming Latency via Thinking-while-Talking</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yangzhuo</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shengpeng</namePart>
<namePart type="family">Ji</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yifu</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tianle</namePart>
<namePart type="family">Liang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haoyu</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name>
<namePart>Junboli</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Fang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lin</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qingyang</namePart>
<namePart type="family">Hong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>Integrating explicit Chain-of-Thought (CoT) into end-to-end spoken dialogue models enhances intelligence but incurs prohibitive latency. While the “Thinking-while-Talking” paradigm alleviates this delay, it fundamentally compromises block atomicity, severing the logical connection between interleaved thought and speech. To address this, we present Dual-Reasoner, employing a Streaming Masking Mechanism underpinned by our Dual-Think-30k dataset to guarantee uninterrupted audio streaming. Crucially, to strictly align the fragmented thinking blocks to service speech generation, we introduce the Atomic-Consistency Restoration framework. To secure comprehensive capabilities in high-difficulty reasoning, this mechanism utilizes a quadruple-constraint system to reconstruct logical atomicity, ensuring that “think” chunks act as a rigorous anchor for “talk” outputs. Experimental results demonstrate that Dual-Reasoner achieves comprehensive reasoning enhancements within ultra-low latency constraints: it elevates the VoiceBench score from 67.24 to 73.41 over the baseline, while significantly reducing the Time-to-First-Audio (TTFA) from 20.35s to 3.65s and the Real-Time Factor (RTF) from 7.04 to 1.05.</abstract>
<identifier type="citekey">li-etal-2026-dual</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.199/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>4081</start>
<end>4105</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Dual-Reasoner: Bridging Interleaved Atomicity and Streaming Latency via Thinking-while-Talking
%A Li, Yangzhuo
%A Ji, Shengpeng
%A Chen, Yifu
%A Liang, Tianle
%A Yang, Haoyu
%A Fang, Jun
%A Li, Lin
%A Hong, Qingyang
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%A Junboli
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F li-etal-2026-dual
%X Integrating explicit Chain-of-Thought (CoT) into end-to-end spoken dialogue models enhances intelligence but incurs prohibitive latency. While the “Thinking-while-Talking” paradigm alleviates this delay, it fundamentally compromises block atomicity, severing the logical connection between interleaved thought and speech. To address this, we present Dual-Reasoner, employing a Streaming Masking Mechanism underpinned by our Dual-Think-30k dataset to guarantee uninterrupted audio streaming. Crucially, to strictly align the fragmented thinking blocks to service speech generation, we introduce the Atomic-Consistency Restoration framework. To secure comprehensive capabilities in high-difficulty reasoning, this mechanism utilizes a quadruple-constraint system to reconstruct logical atomicity, ensuring that “think” chunks act as a rigorous anchor for “talk” outputs. Experimental results demonstrate that Dual-Reasoner achieves comprehensive reasoning enhancements within ultra-low latency constraints: it elevates the VoiceBench score from 67.24 to 73.41 over the baseline, while significantly reducing the Time-to-First-Audio (TTFA) from 20.35s to 3.65s and the Real-Time Factor (RTF) from 7.04 to 1.05.
%U https://aclanthology.org/2026.findings-acl.199/
%P 4081-4105
Markdown (Informal)
[Dual-Reasoner: Bridging Interleaved Atomicity and Streaming Latency via Thinking-while-Talking](https://aclanthology.org/2026.findings-acl.199/) (Li et al., Findings 2026)
ACL
- Yangzhuo Li, Shengpeng Ji, Yifu Chen, Tianle Liang, Haoyu Yang, Junboli, Jun Fang, Lin Li, and Qingyang Hong. 2026. Dual-Reasoner: Bridging Interleaved Atomicity and Streaming Latency via Thinking-while-Talking. In Findings of the Association for Computational Linguistics: ACL 2026, pages 4081–4105, San Diego, California, United States. Association for Computational Linguistics.