Rating is available when the video has been rented.
This feature is not available right now. Please try again later.
Published on Feb 17, 2015
Jonathan Dinu - Distributed Machine Learning: Architectures to Leverage Streams of Data
With the ever increasing amount of data produced by IoT devices, rich experiences are possible that before were only the realm of science fiction. From data driven applications such as Google Now, to smart home automation devices like Nest, we are seeing companies leverage machine learning to deliver personalized products. But with the demand of users for seamless experiences engineers must be able to deliver results seemingly effortlessly and in realtime.
In this talk I will cover different approaches to deploying a production machine learning application on massive amounts of data to deliver insight in near-realtime. By leveraging distributed architectures as well as streaming and online algorithms you can intelligently break up a larger problem to create a responsive data product. But with any approach (streaming or batch) there are always tradeoffs in how you design your system and knowing each's limits is critical.