Scholarlink

In the wake of Industry 4.0’s progression, General Artificial Intelligence (AGI) systems, such as OpenNARS, have emerged as pivotal in both augmenting and potentially supplanting domain-specific AI in industrial contexts. Since AGI is not inherently designed for industrial tasks, it is necessary to tailor them from general-purpose to specialized applications to address their efficiency shortcomings. This paper, based on the transformation of AGI’s priors design and educational guidance undertakes a specialized optimization of OpenNARS for an industrial scenario using the real-world case of steel surface defect detection. The findings demonstrate that the resultant specialized OpenNARS system manifests a marked enhancement in both reasoning efficiency and accuracy, which is a significant advancement for the broader adoption of AGI in industrial settings.

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

The Specialization of AGI: Exploration of Industrial Applications for General Artificial Intelligence

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Tao Zhang,
  • Changxin Sun,
  • Tao Wang,
  • Zhenyu Wu,
  • Ketao Mo,
  • Weilin Shen,
  • Pei Wang,
  • Kai Liu

摘要

In the wake of Industry 4.0’s progression, General Artificial Intelligence (AGI) systems, such as OpenNARS, have emerged as pivotal in both augmenting and potentially supplanting domain-specific AI in industrial contexts. Since AGI is not inherently designed for industrial tasks, it is necessary to tailor them from general-purpose to specialized applications to address their efficiency shortcomings. This paper, based on the transformation of AGI’s priors design and educational guidance undertakes a specialized optimization of OpenNARS for an industrial scenario using the real-world case of steel surface defect detection. The findings demonstrate that the resultant specialized OpenNARS system manifests a marked enhancement in both reasoning efficiency and accuracy, which is a significant advancement for the broader adoption of AGI in industrial settings.