@inproceedings{lyu-etal-2026-touchthinker,
title = "{T}ouch{T}hinker: Scaling Tactile Commonsense Reasoning to the Open World with Large-scale Data and Action-aware Representation",
author = "Lyu, Kailin and
Wu, Di and
Zhang, Pengwei and
Zheng, Yuhang and
Lai, Yingxin and
Xiao, Long and
Wu, Kangyi and
Li, Pengna and
Gao, Chen and
Hu, Lianyu and
Hu, Xiaobin and
Hao, Jie and
Hao, Ce and
Yuan, Weihao and
Yan, Shuicheng",
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.1925/",
pages = "43841--43858",
ISBN = "979-8-89176-459-0",
abstract = "Touch is a key modality for embodied agents to understand the physical world. Although recent work has incorporated tactile signals into lan- guage systems for tactile Commonsense Rea- soning, scaling such systems to realistic open- world settings remains challenging due to two key bottlenecks: (1) current tactile reasoning datasets remain limited in format and scale, pro- viding insufficient supervision for reasoning from tactile observations to physical common- sense and hindering the learning of transferable tactile commonsense; (2) Tactile signals are inherently redundant and action-specific, yet existing methods often overlook these proper- ties, resulting in inefficient representations with limited semantic expressiveness. To address these limitations, we propose TouchThinker, a tactile-language framework that scales tac- tile commonsense reasoning to the open world from both data and representation perspec- tives. First, we construct TouchThinker-1M, a million-scale, multi-source tactile reasoning dataset covering 410 objects, 8 scenarios, and 7 sensor types, providing a solid data founda- tion for open-world generalization. We fur- ther introduce TouchThinker-Bench, an open- world benchmark with more realistic and di- verse tasks. Then, we further introduce action- aware tactile modeling and a two-stage training paradigm to improve tactile representation effi- ciency and enable efficient reasoning. Experi- mental results demonstrate that TouchThinker achieves competitive performance against state- of-the-art models across multiple datasets. Our code and dataset will be released."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lyu-etal-2026-touchthinker">
<titleInfo>
<title>TouchThinker: Scaling Tactile Commonsense Reasoning to the Open World with Large-scale Data and Action-aware Representation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kailin</namePart>
<namePart type="family">Lyu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Di</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pengwei</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuhang</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yingxin</namePart>
<namePart type="family">Lai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Long</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kangyi</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pengna</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chen</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lianyu</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaobin</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jie</namePart>
<namePart type="family">Hao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ce</namePart>
<namePart type="family">Hao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weihao</namePart>
<namePart type="family">Yuan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shuicheng</namePart>
<namePart type="family">Yan</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>Touch is a key modality for embodied agents to understand the physical world. Although recent work has incorporated tactile signals into lan- guage systems for tactile Commonsense Rea- soning, scaling such systems to realistic open- world settings remains challenging due to two key bottlenecks: (1) current tactile reasoning datasets remain limited in format and scale, pro- viding insufficient supervision for reasoning from tactile observations to physical common- sense and hindering the learning of transferable tactile commonsense; (2) Tactile signals are inherently redundant and action-specific, yet existing methods often overlook these proper- ties, resulting in inefficient representations with limited semantic expressiveness. To address these limitations, we propose TouchThinker, a tactile-language framework that scales tac- tile commonsense reasoning to the open world from both data and representation perspec- tives. First, we construct TouchThinker-1M, a million-scale, multi-source tactile reasoning dataset covering 410 objects, 8 scenarios, and 7 sensor types, providing a solid data founda- tion for open-world generalization. We fur- ther introduce TouchThinker-Bench, an open- world benchmark with more realistic and di- verse tasks. Then, we further introduce action- aware tactile modeling and a two-stage training paradigm to improve tactile representation effi- ciency and enable efficient reasoning. Experi- mental results demonstrate that TouchThinker achieves competitive performance against state- of-the-art models across multiple datasets. Our code and dataset will be released.</abstract>
<identifier type="citekey">lyu-etal-2026-touchthinker</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1925/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>43841</start>
<end>43858</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T TouchThinker: Scaling Tactile Commonsense Reasoning to the Open World with Large-scale Data and Action-aware Representation
%A Lyu, Kailin
%A Wu, Di
%A Zhang, Pengwei
%A Zheng, Yuhang
%A Lai, Yingxin
%A Xiao, Long
%A Wu, Kangyi
%A Li, Pengna
%A Gao, Chen
%A Hu, Lianyu
%A Hu, Xiaobin
%A Hao, Jie
%A Hao, Ce
%A Yuan, Weihao
%A Yan, Shuicheng
%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 lyu-etal-2026-touchthinker
%X Touch is a key modality for embodied agents to understand the physical world. Although recent work has incorporated tactile signals into lan- guage systems for tactile Commonsense Rea- soning, scaling such systems to realistic open- world settings remains challenging due to two key bottlenecks: (1) current tactile reasoning datasets remain limited in format and scale, pro- viding insufficient supervision for reasoning from tactile observations to physical common- sense and hindering the learning of transferable tactile commonsense; (2) Tactile signals are inherently redundant and action-specific, yet existing methods often overlook these proper- ties, resulting in inefficient representations with limited semantic expressiveness. To address these limitations, we propose TouchThinker, a tactile-language framework that scales tac- tile commonsense reasoning to the open world from both data and representation perspec- tives. First, we construct TouchThinker-1M, a million-scale, multi-source tactile reasoning dataset covering 410 objects, 8 scenarios, and 7 sensor types, providing a solid data founda- tion for open-world generalization. We fur- ther introduce TouchThinker-Bench, an open- world benchmark with more realistic and di- verse tasks. Then, we further introduce action- aware tactile modeling and a two-stage training paradigm to improve tactile representation effi- ciency and enable efficient reasoning. Experi- mental results demonstrate that TouchThinker achieves competitive performance against state- of-the-art models across multiple datasets. Our code and dataset will be released.
%U https://aclanthology.org/2026.emnlp-main.1925/
%P 43841-43858
Markdown (Informal)
[TouchThinker: Scaling Tactile Commonsense Reasoning to the Open World with Large-scale Data and Action-aware Representation](https://aclanthology.org/2026.emnlp-main.1925/) (Lyu et al., EMNLP 2026)
ACL
- Kailin Lyu, Di Wu, Pengwei Zhang, Yuhang Zheng, Yingxin Lai, Long Xiao, Kangyi Wu, Pengna Li, Chen Gao, Lianyu Hu, Xiaobin Hu, Jie Hao, Ce Hao, Weihao Yuan, and Shuicheng Yan. 2026. TouchThinker: Scaling Tactile Commonsense Reasoning to the Open World with Large-scale Data and Action-aware Representation. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 43841–43858, Budapest, Hungary. Association for Computational Linguistics.