During the last decade, new machine learning (ML) methods have been developed to solve complex tasks with an impressive performance that were until recently believed to be impossible. However, many of these approaches may fail in many networking scenarios, where limited computational and communication resources prohibit heavy iterative algorithms, privacy concerns hinder sharing the datasets, and the underlying communication graph enforces inference with partial knowledge. Multi-agent systems, Internet-of-Things, transportation networks, engineered biological networks, social networks, and intra-body sensor networks are examples of such networks, where the traditional ML approaches may fail. This workshop will focus on pioneering works targeting machine learning over networks (MLoNs). Topics of interest include, but not limited to, the following:

• Model compression and efficient distributed ML
• Compressed gradient methods and error compensation
• Distributed learning on non-IID datasets
• Federated learning and privacy-preserving distributed ML

The workshop will feature some invited talks, student presentations, and a panel on the mentioned topics.