Machine learning provides the foundation for many smart applications. Therefore, a large amount of partly private data is captured by sensors, pre-processed by data stream systems and stored in databases. Such applications are a substantial benefit for the user. Yet, there is a growing concern on the part of users regarding the large-scale processing of private data. In addition, new regulations such as the General Data Protection Regulation (GDPR) restrict this kind of data processing even further. The PriMaL Special Track concerns with novel approaches that guarantee privacy in machine learning applications without restricting their utility unnecessarily.