文献类型：期刊文献

英文题名：Enhancing intention understanding in voice assistants for older adults using large language models

作者：Fu, Jiaojiao[1];Kang, Chun[2];Qian, Zhigu[2];Zhou, Yangfan[2]

机构：[1]East China Univ Sci & Technol, Sch Informat Sci & Engn, Shanghai, Peoples R China;[2]Fudan Univ, Shanghai Key Lab Intelligent Informat Proc, Sch Comp Sci, Shanghai, Peoples R China

年份：2026

卷号：45

期号：13

起止页码：3285

外文期刊名：BEHAVIOUR & INFORMATION TECHNOLOGY

收录：;EI(收录号：20260820106464);WOS:【SSCI(收录号:WOS:001689495800001)，SCI-EXPANDED(收录号:WOS:001853377100001)】；

基金：We would like to thank all the participants for taking part in our study and making this research possible. This work is supported by Shanghai Key Laboratory of Intelligent Information Processing, Fudan University (Grant No. IIPL-2025-RD4-04).

语种：英文

外文关键词：Older adults; voice assistants; intention understanding; smartphone operation; large language models

摘要：Voice assistants offer potential for aiding older adults with mobile app navigation, yet understanding their intentions remains challenging. To address this, we explore using large language models to enhance intention comprehension in voice assistants for older adults. Qualitative research on Chinese older adults reveals five categories of expression failures: localised expressions, colloquialisms, multitasking, incomplete expressions, and ambiguous expressions. We propose tailored methods for each, employing prompts to translate original instructions into formal ones and interpret free text instructions into structured apps and operations. Experiments with 44 tasks reflecting real interaction failures validate our approach's effectiveness. By tackling intention understanding failure, our work improves voice assistant usability for older adults, contributing to AI technologies' inclusive advancement.