A research-led newsletter from an optical communications group that explains how interpretable machine learning can yield analytical insight into nonlinear distortions in fiber-optic transmission and improve digital signal processing. It reports on experimental testbed results and frames ML as a tool to complement physics-based analysis.