Embedded Machine Learning

SEA Lab has a strong commitment to developing resource-constrained embedded systems that can support machine learning methods. This research line focuses on the applications that may benefit from data-driven models but at the same time set hard requirements for power consumption, size, and cost. 

The research activities can be arranged within the following macro categories:

  • computer vision with embedded machine learning for robotics and prosthetics
  • digital architectures for efficient implementation of single layer feed forward neural networks on FPGAs
  • machine learning models for Internet of Things

Please check some videos that demonstrate the operation of the techniques and methods developed in this line of research about Deep Learning in embedded systems:

 

Last update 8 November 2022