Scholarlink

<p>Public health crises challenge information systems with rampant misinformation. This can trigger a “sensemaking collapse,” undermining organizational response and public trust. Traditional fake news detection methods, often decontextualized, are inadequate for these dynamic, knowledge-intensive events. This paper introduces EKAN (External Knowledge-Augmented Attention Neural Network), a novel framework grounded in situated cognition. EKAN functions as a computational augmentation tool. It integrates dynamic external knowledge from authoritative sources with internal content features via an attention mechanism. This process creates a “cognitive anchor” that enables context-aware veracity assessment. Validated on the Ohio Train Derailment dataset, EKAN significantly outperforms static baselines. This highlights the criticality of dynamic knowledge contextualization for robust countermeasures. It also offers a pathway toward designing more epistemically resilient crisis information systems that can support effective sensemaking amidst informational uncertainty.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

EKAN: a situated cognition framework integrating dynamic knowledge and attention for fake news countermeasures in crisis information systems

  • 期刊论文

  • 发表日期：

    2026年6月15日

查看全文
  • Jianhong Luo,
  • Aodong Hao

摘要

Public health crises challenge information systems with rampant misinformation. This can trigger a “sensemaking collapse,” undermining organizational response and public trust. Traditional fake news detection methods, often decontextualized, are inadequate for these dynamic, knowledge-intensive events. This paper introduces EKAN (External Knowledge-Augmented Attention Neural Network), a novel framework grounded in situated cognition. EKAN functions as a computational augmentation tool. It integrates dynamic external knowledge from authoritative sources with internal content features via an attention mechanism. This process creates a “cognitive anchor” that enables context-aware veracity assessment. Validated on the Ohio Train Derailment dataset, EKAN significantly outperforms static baselines. This highlights the criticality of dynamic knowledge contextualization for robust countermeasures. It also offers a pathway toward designing more epistemically resilient crisis information systems that can support effective sensemaking amidst informational uncertainty.