With the rapid and recent rise of data science, machine learning frameworks such as TensorFlow have become popular. However, those frameworks do not form a complete Machine Learning Platform by themselves. In this talk we will see what role Databases play in the Machine Learning World, in particular Multi-Model databases supporting multiple data models such as graphs, documents, and key-values.
Many powerful Machine Learning algorithms are based on graphs, e.g., Page Rank (Pregel), Recommendation Engines (collaborative filtering), text summarization and other NLP tasks.
There are even more applications once we consider data pre-processing and feature engineering which are both vital tasks in Machine Learning Pipelines.
But how can we combine Multi-Model Databases with Machine Learning Systems such as TensorFlow or Pytorch?
Using real-world examples we show how Multi-Model databases and machine learning frameworks form a very powerful combination. In particular, we will focus on graph-based Machine Learning models, as well as, graph-based data pre-processing and feature engineering (which can, in turn, serve as input for a deep neural network).