Training strategies for learned Volterra MIMO equalizers in WDM systems

This research advances training strategies for a learned Volterra MIMO equaliser tailored for Wavelength Division Multiplexing (WDM) systems. Building upon prior success with minimal steps per span, our current focus is improving training efficiency. The equaliser integrates filters to address key signal impairments, and our work delves into refining training strategies for these filters. As part of the considered techniques, we explore the potential of applying transfer learning among the channels in the equaliser. This investigation yields valuable insights into the interplay of training strategies and filter optimisation, contributing to the enhanced performance of learned Volterra MIMO equalisers in WDM systems.