Outlier Detection at Netflix

Re:Outlier Detection at Netflix by cbsmith on Tuesday July 14, 2015 @05:47PM (#9888296)

There are a lot of online methods for reinforcement learning, and one could argue that reinforcement learning is doing exactly what you want for outlier detection: doing a cost benefit analysis of the value of exploring a given context in more or less detail.

Have you guys thought at all about adapting a reinforcement learning model for building an online model to detect outliers?

[ Parent | on HN ]

Re:Outlier Detection at Netflix by diab0lic on Tuesday July 14, 2015 @06:07PM (#9888400)

We have, specifically in the domain of using RL for parameter selection. Part of the reason I was brought on board was background experience in RL. I fully believe it's the key to turning domain (yes/no) decisions into numbers for our algorithms to ingest.

Ultimately we would like all of our systems to use RL to interpret user feedback to determine parameters, actions, etc...

Brilliant question by the way!

[ Parent | on HN ]

Re:Outlier Detection at Netflix by cbsmith on Wednesday July 15, 2015 @03:09AM (#9890107)

Brilliant answer. :-)

Sounds like you guys have thought this through really well.

[ Parent | on HN ]