Apache SystemML is a declarative style language designed for large-scale machine learning. It provides automatic generation of optimized runtime plans ranging from single-node, to in-memory, to distributed computations on Apache Hadoop and Apache Spark. SystemML algorithms are expressed in R-like or Python-like syntax that includes linear algebra primitives, statistical functions and ML-specific constructs. As a data scientist, engineer, or just a fellow interested in machine learning, your productivity will increase while having the flexibility to express custom analytics and not worry about the underlying optimization engine. Automatic scalability and optimization is handled by SystemML. This course will not only provide you with a view of how the optimizers function but also provide hands-on examples of ML algorithms and how to run them.
Overview
Syllabus
- Module 1 - What is SystemML?
- Explain the purpose and the origin of SystemML
- List the alternatives to SystemML
- Compare performances of SystemML with the alternatives
- Module 2 - SystemML and the Spark MLContext
- Use MLContext to interact with SystemML (in Scala)
- Module 3 - Working with BigSheets
- Describe and use a number of SystemML algorithms
- Module 4 - Working with BigSheets
- Explain the purpose of DML
- Describe the DML language
- List some of the built-in functions
- Module 5 - Working with BigSheets
- Describing the optimizer stack
- Explaining how SystemML know it's better to run on one machine
- Explaining why SystemML is so much faster than single-node R
