Machine Learning for Energy-efficient, Reliable Manycore Interconnects

Track Chairs

Amey Kulkrani (NVIDIA, USA)

Topic of Interest

Computing systems have been an essential factor in the advancement of technology and applications, and machine learning is no different. Large servers have enabled complex machine learning algorithms to achieve high performance in cognitive applications. However, with the rising need for advanced machine learning for large-scale applications, scalable, low-cost, high-performance, and energy-efficient manycore systems on chip (MCSoCs) have become crucial. The topics of interest of the track include, but are not limited to:

  • ML for multicore SoC Design
  • Thermal Design Optimization for 3D-IC
  • ML for Hardware Design
  • IP-obscuring models for non-linear circuits
  • Model-based design with surrogate models
  • Bayesian optimization to find the region of design space