In-Database Factorised Learning - Professor Dan Olteanu

Explore in-database factorised learning with Professor Dan Olteanu in this one-hour seminar from the Alan Turing Institute. Delve into advanced database concepts, including relational and XML query processing, incomplete information, probabilistic databases, and factorised databases. Learn about the evolution of database integration, from loose to tight integration, and understand the advantages of unified programming architectures. Examine practical applications such as linear regression, encoding categorical variables, and optimizing queries. Discover how conditional independence and factorizations can improve database performance. Investigate functional dependencies, penalty terms, and polynomial regression in the context of factorised learning. Gain insights into feature extraction techniques and their impact on database efficiency.

Intro
Overview
No Integration
Loose Integration
Unified Programming Architecture
Tight Integration
Advantages
Schematic
Results
Target
Database
Linear regression
Encoding categorical variables
Alpha
First order query
Categorical query
In the ultimate
Optimally
Example
Conditional Independence
Factorizations
Recap
Functional dependency
Function for country
Function for city
Penalty term
Conclusion
polynomial regression
factorisation
feature extraction