Scholarlink

In this paper the Deep Unfolding is presented as an approach to build complex scientific computing algorithms with an online cost affordable by embedded digital twins. We give two specific algorithmic examples: Nonnegative Matrix Factorization and the Kalman Filter. For both of them it is also presented a solution approach viable on embedded digital twins for audio source separation and for thermographic inspections.

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

Exploiting Scientific Machine Learning on Embedded Digital Twins

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Erik Chinellato,
  • Paolo Martin,
  • Laura Rinaldi,
  • Fabio Marcuzzi

摘要

In this paper the Deep Unfolding is presented as an approach to build complex scientific computing algorithms with an online cost affordable by embedded digital twins. We give two specific algorithmic examples: Nonnegative Matrix Factorization and the Kalman Filter. For both of them it is also presented a solution approach viable on embedded digital twins for audio source separation and for thermographic inspections.