ML4DM: The second workshop on the emerging applications of machine learning in modern data management.

The classic approaches to solving various problems in data management either rely on heuristics and assumptions that often do not hold in practice or estimates based on crude statistics. These limitations have motivated the adoption of ML-based methods at a pace not seen before. These new methods not only reduce the cost of development and maintenance but are automatically tailored to fit individual system instances. Inspired by these advancements, the workshop speakers presented examples of using machine learning for modernizing different aspects of data management in the areas of Query Optimization, Data Partitioning, Database Knobs Tuning, and Workload Management.