Online Machine Learning With Distributed In-memory Clusters

Data Science
Location: King's Suite - Sandringham Level: Intermediate
Average rating: ***..
(3.57, 7 ratings)
Slides:   1-PPT 
Traditionally machine learning algorithms are retrained periodically in a batch fashion. A champion/challenger or manual assessment approach is used to determine if a production model requires an updated.

An increasing number of use cases in risk, intrastructure security and other areas are benefiting from a streaming approach to model retraining using each new observation as an opportunity to adjust the algorithm. The technique is being used for algorithms that perform Classification, Regression, Recommendation, Graph Mining and Anomaly Detection.

The availability of inexpensive in-memory distributed clustering technilogies is allowing the online machine learning approach.

Photo of Arshak Navruzyan

Arshak Navruzyan

Argyle Data

Arshak is a passionate technologist with a background in distributed systems and machine learning. He currently serves as VP Product Management at Argyle Data, focused on a complete platform for data-driven applications.

Previously Arshak was VP Product Management of Alpine Data Labs and has held senior technical and product management roles at Endeca and Oracle. Arshak graduated from UCLA.

Sponsors

Sponsorship Opportunities

For exhibition and sponsorship opportunities, contact Susan Stewart at sstewart@oreilly.com

Media Partner Opportunities

For information on trade opportunities with O'Reilly conferences email mediapartners
@oreilly.com

Press & Media

For media-related inquiries, contact Maureen Jennings at maureen@oreilly.com

Contact Us

View a complete list of Strata contacts