17:50 | 18:20
Keywords defining the session:
- Machine Learning
Takeaway points of the session:
- Notebooks provide the agility to interactively execute analysis of different nature (e.g., graph-analytics, statistics, ML) across various programming languages (e.g., R, python, groovy, SQL).
Notebooks have been emerging as a key tool for handling big data: they provide agility in data analysis and prototyping and bring together a wide range of techniques and languages.
One key synergy that a notebook environment enables is between graph analytics and machine learning.
Connections in the data often hold key information that is not immediately available to classic machine learning techniques, but can be surfaced via graph analytics tools.
Graph-empowered machine learning techniques (e.g., Node2vec, Deepwalk or Graph Convolutional Neural Networks) exploit this insight and find a natural environment in notebooks, which provide agility of prototyping, a way for explainability through visualization, and performance, with the ability to bring together in subsequent paragraphs different, specialized execution engines.