Categories
Machine Learning

Fascinating paper by Microsoft researchers, bringing the entire Machine Learning lifecycle to DBMS, after NoSQL, is MLSQL going to be next?

I read this morning paper twice, to let the idea sink in. The acronym EGML isn’t that exciting though, it sounds too close to SGML and HTML for comfort. As I read it I couldn’t help notice the stark contrast to the approach taken by AWS with their machine learning offering.

This paper describes EGML, clearly at an exploratory stage, it raises some interesting thought provoking questions. If the DBMS were to become home to entreprise grade ML, wouldn’t that add entire new dimensions to SQL, what kind of mathematical foundation should that be built on? It’s too early to tell whether this would eventually result in a specific product, but many of the concepts are likely to live on in some form.

An excerpt from Adrian’s Morning Paper:

Cloudy with a high chance of DBMS: a 10-year prediction for enterprise-grade ML, Agrawal et al., CIDR’20 “Cloudy with a high chance of DBMS” is a fascinating vision paper from a group of experts at Microsoft, looking at the transition of machine learning from being primarily the domain of large-scale, high-volume consumer applications to being…

— Read on blog.acolyer.org/2020/02/19/ten-year-egml-predictions/

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.