@inproceedings{lee-etal-2026-hear,
title = "{HEAR} Who Said What: Unlocking Speaker-Attributed Reasoning via Counterfactual Voice Grounding",
author = "Lee, Dongwook and
Park, Sangkwon and
Song, Eunwoo and
Lee, Che Hyun and
Cho, Youngho and
Kim, Junho and
Yi, June Young and
Kim, Heeseung and
Yoon, Sungroh",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2503/",
pages = "56905--56932",
ISBN = "979-8-89176-459-0",
abstract = "Speech Language Models (SLMs) are increasingly deployed in multi-speaker environments, yet their ability to attribute speech to the correct speaker and reason over speaker identities remains unclear. Hence, we introduce **HEAR**, a conceptually hierarchical benchmark diagnosing the foundational capabilities of speaker-attributed reasoning, comprising 2.4K human-verified samples from 887 diverse multi-party audio clips. Evaluating 20 leading SLMs on HEAR reveals they struggle with these foundational tasks, often relying on semantic priors rather than actual vocal cues. To address this, we present **A2R**, a 30B model optimized on *Counterfactual Audio with Speaker-level Hard negatives* (**CASH**), a dataset designed to guide the model to prioritize acoustic vocal cues over linguistic signals. A2R achieves strong performance on HEAR and exhibits zero-shot generalization to diverse multi-speaker downstream tasks, demonstrating that learned speaker attribution unlocks the model{'}s latent capacity for speaker-aware reasoning. All resources are available on our https://attributetoreason.github.io/AttributeToReason/"
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lee-etal-2026-hear">
<titleInfo>
<title>HEAR Who Said What: Unlocking Speaker-Attributed Reasoning via Counterfactual Voice Grounding</title>
</titleInfo>
<name type="personal">
<namePart type="given">Dongwook</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sangkwon</namePart>
<namePart type="family">Park</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eunwoo</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Che</namePart>
<namePart type="given">Hyun</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Youngho</namePart>
<namePart type="family">Cho</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junho</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">June</namePart>
<namePart type="given">Young</namePart>
<namePart type="family">Yi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Heeseung</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sungroh</namePart>
<namePart type="family">Yoon</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Speech Language Models (SLMs) are increasingly deployed in multi-speaker environments, yet their ability to attribute speech to the correct speaker and reason over speaker identities remains unclear. Hence, we introduce **HEAR**, a conceptually hierarchical benchmark diagnosing the foundational capabilities of speaker-attributed reasoning, comprising 2.4K human-verified samples from 887 diverse multi-party audio clips. Evaluating 20 leading SLMs on HEAR reveals they struggle with these foundational tasks, often relying on semantic priors rather than actual vocal cues. To address this, we present **A2R**, a 30B model optimized on *Counterfactual Audio with Speaker-level Hard negatives* (**CASH**), a dataset designed to guide the model to prioritize acoustic vocal cues over linguistic signals. A2R achieves strong performance on HEAR and exhibits zero-shot generalization to diverse multi-speaker downstream tasks, demonstrating that learned speaker attribution unlocks the model’s latent capacity for speaker-aware reasoning. All resources are available on our https://attributetoreason.github.io/AttributeToReason/</abstract>
<identifier type="citekey">lee-etal-2026-hear</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2503/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>56905</start>
<end>56932</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T HEAR Who Said What: Unlocking Speaker-Attributed Reasoning via Counterfactual Voice Grounding
%A Lee, Dongwook
%A Park, Sangkwon
%A Song, Eunwoo
%A Lee, Che Hyun
%A Cho, Youngho
%A Kim, Junho
%A Yi, June Young
%A Kim, Heeseung
%A Yoon, Sungroh
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F lee-etal-2026-hear
%X Speech Language Models (SLMs) are increasingly deployed in multi-speaker environments, yet their ability to attribute speech to the correct speaker and reason over speaker identities remains unclear. Hence, we introduce **HEAR**, a conceptually hierarchical benchmark diagnosing the foundational capabilities of speaker-attributed reasoning, comprising 2.4K human-verified samples from 887 diverse multi-party audio clips. Evaluating 20 leading SLMs on HEAR reveals they struggle with these foundational tasks, often relying on semantic priors rather than actual vocal cues. To address this, we present **A2R**, a 30B model optimized on *Counterfactual Audio with Speaker-level Hard negatives* (**CASH**), a dataset designed to guide the model to prioritize acoustic vocal cues over linguistic signals. A2R achieves strong performance on HEAR and exhibits zero-shot generalization to diverse multi-speaker downstream tasks, demonstrating that learned speaker attribution unlocks the model’s latent capacity for speaker-aware reasoning. All resources are available on our https://attributetoreason.github.io/AttributeToReason/
%U https://aclanthology.org/2026.emnlp-main.2503/
%P 56905-56932
Markdown (Informal)
[HEAR Who Said What: Unlocking Speaker-Attributed Reasoning via Counterfactual Voice Grounding](https://aclanthology.org/2026.emnlp-main.2503/) (Lee et al., EMNLP 2026)
ACL
- Dongwook Lee, Sangkwon Park, Eunwoo Song, Che Hyun Lee, Youngho Cho, Junho Kim, June Young Yi, Heeseung Kim, and Sungroh Yoon. 2026. HEAR Who Said What: Unlocking Speaker-Attributed Reasoning via Counterfactual Voice Grounding. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 56905–56932, Budapest, Hungary. Association for Computational Linguistics.