Scholarlink

The IoT's rapid expansion into smart homes and industrial automation has greatly increased device connectivity, driving widespread adoption. However, IoT systems face growing security risks like flooding, intrusion, and spoofing attacks. Traditional defenses such as firewalls and IDS struggle with the complexity and diversity of these threats, compounded by overlapping classes in IoT data. To address these challenges, this study proposes a deep learning-based IDS integrating Stacked Autoencoders (SAE) for nonlinear dimensionality reduction, resolving class overlap issues while preserving crucial features. Long Short-Term Memory networks (LSTM) are utilized for their temporal data processing capabilities to identify various attack types. This research aims to develop a precise and robust IoT security mechanism, enhancing attack detection accuracy and bolstering system security in practical applications. It lays a foundation for future research in this critical area.

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

Multi-class Attack Detection in IoT Using Dimensionality Reduction with Stacked Autoencoders and Classification with Long Short-Term Memory Networks

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yu-Xian Lee,
  • Chin-Shiuh Shieh,
  • Mong-Fong Horng,
  • Ying-Chieh Chao,
  • Prasun Chakrabarti

摘要

The IoT's rapid expansion into smart homes and industrial automation has greatly increased device connectivity, driving widespread adoption. However, IoT systems face growing security risks like flooding, intrusion, and spoofing attacks. Traditional defenses such as firewalls and IDS struggle with the complexity and diversity of these threats, compounded by overlapping classes in IoT data. To address these challenges, this study proposes a deep learning-based IDS integrating Stacked Autoencoders (SAE) for nonlinear dimensionality reduction, resolving class overlap issues while preserving crucial features. Long Short-Term Memory networks (LSTM) are utilized for their temporal data processing capabilities to identify various attack types. This research aims to develop a precise and robust IoT security mechanism, enhancing attack detection accuracy and bolstering system security in practical applications. It lays a foundation for future research in this critical area.