After a delay due to the COVID-19 pandemic, the third Statistical Learning for Signal and Image Processing Workshop 2021 (SLSIP) finally took place in Rüdesheim in October. After the success of two previous editions in Helsinki 2018 and Annecy 2019, the third edition of the workshop focused on methodological aspects and application challenges: The main topics concerned machine learning for robust signal processing and control, biomedical signal and image processing, statistical modelling of sparse, complex and multi-sensor data, graph theory for signal processing in dynamic and large-scale networks, and distributed optimization for smart systems with applications, such as, neuroscience, electric systems, array signal processing. “The workshop was a great success with outstanding feedback from the 40 participants,” summarised emergenCITY researcher Michael Muma afterwards. Another highlight of the three days was the emergenCITY-keynote by David Tyleron “Robustness properties of regularized M-estimates of covariance: A statistician’s view”. David E. Tyler, Distinguished Professor at Rutgers University, USA, is a pioneer in the field of robust statistics.