EUVIMEDEUVIMED entdecken
Crossref

Enhanced Stance Detection via Knowledge Retrieval and Contextual Integration in Large Language Models

Shangfei Wei, Zhaohong Lai, Biao Fu, Xiaodong Shi

IEEE Sensors Journal · 01.07.2026 · 26 (13) · 19912-19922

DOI: 10.1109/jsen.2026.3689656 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Sensors Journal, erschienen am 01.07.2026. Verfasst von Shangfei Wei und 3 weiteren Autor:innen. Die Arbeit nennt 38 Literaturangaben. Gefördert von: National Key Research and Development Program of China, Key Support Project of NSFC-Liaoning Joint Foundation, Major Scientific Research Project of the State Language Commission in the 13th Five-Year Plan.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

With the popularity of the Internet and mobile devices, communication and stance expression between users are increasingly carried out through various social media platforms. Social sensing refers to the collection, processing, and analysis of data from social media and Internet communications. This new form of perception mode can provide a deeper understanding of users’ emotions and stance expressions. Stance detection is a crucial task in natural language processing (NLP) and involves identifying the stance concerning a specific target in a text. Traditional approaches primarily focused on explicit stance expressions; however, the rise of brief and often ambiguous content in online social media introduces significant challenges. Recently, large language models (LLMs) have demonstrated strong generalization capabilities across various NLP tasks, including stance detection. However, LLMs may generate biased or erroneous stance predictions. In this article, we propose a novel framework, knowledge retrieval and contextual integration stance detection (KRCISD), that enhances stance detection by integrating effective background knowledge and enlightening viewpoints generated through LLMs. We retrieve relevant background information from external knowledge sources and use this data to guide LLMs in generating context-aware perspectives that the strategy can effectively enhance LLMs’ ability to discern stance. We validate the effectiveness of the KRCISD framework through extensive experiments across multiple models and datasets, demonstrating its ability to enhance stance detection performance significantly.

Förderung

  • National Key Research and Development Program of China
  • Key Support Project of NSFC-Liaoning Joint Foundation
  • Major Scientific Research Project of the State Language Commission in the 13th Five-Year Plan

Zitieren

APA

Wei, S., Lai, Z., Fu, B., & Shi, X. (2026). Enhanced Stance Detection via Knowledge Retrieval and Contextual Integration in Large Language Models. IEEE Sensors Journal, 26(13), 19912-19922. https://doi.org/10.1109/jsen.2026.3689656

Vancouver

Wei S, Lai Z, Fu B, Shi X. Enhanced Stance Detection via Knowledge Retrieval and Contextual Integration in Large Language Models. IEEE Sensors Journal. 2026;26(13):19912-19922. doi:10.1109/jsen.2026.3689656

BibTeX
@article{Wei2026,
  author = {Wei, Shangfei and Lai, Zhaohong and Fu, Biao and Shi, Xiaodong},
  title = {Enhanced Stance Detection via Knowledge Retrieval and Contextual Integration in Large Language Models},
  journal = {IEEE Sensors Journal},
  year = {2026},
  volume = {26},
  number = {13},
  pages = {19912--19922},
  doi = {10.1109/jsen.2026.3689656}
}
Kostenlos starten →