Presented By O’Reilly and Cloudera
Make Data Work
21–22 May 2018: Training
22–24 May 2018: Tutorials & Conference
London, UK

Learning how to design automatically updating AI with Apache Kafka and DeepLearning4J

Jason Bell (MastodonC)
16:3517:15 Thursday, 24 May 2018

Who is this presentation for?

Data engineers, software developers and IT analysts.

Prerequisite knowledge

Basic programming knowledge is helpful. Though I will offer explanations of how things work as I present the talk.

What you'll learn

Learn the core elements of designing a self learning prediction system using streaming tools and artificial intelligence. Also when to give consideration on when to retrain the system and when human interaction may be required.

Description

With the increased development and adoption of streaming platforms we now have a solid mechanism for collecting and processing data in a timely fashion. Alongside that the growth and interest in machine learning and artificial intelligence has given us refined prediction and decision making.

In this session you will learn how to design a streaming application with continual learning. Jason Bell will use Kafka and DeepLearning4J to illustrate the design and implementation of a system that will accept data, apply training to a neural network and also output predictions. Jason will then look at the considerations that have to be made on how this application can continually learn, when manual intervention is required and how to evaluate incremental learning.

You will learn how to:

  • Plan the application out on paper.
  • Consider which topics are required.
  • Use Kafka Connect to store raw streaming data.
  • Define a DeepLearning4J neural network.
  • Reapply neural network training with new training data.
  • Make predictions using the Kafka Streaming API

This talk is intended for anyone with an interest in the applications that machine learning and deep learning have in an increasingly streamed world. While the focus of the talk is on the open source tools available the techniques learned from this talk can be applied to other learning and streaming platforms.

Photo of Jason Bell

Jason Bell

MastodonC

Jason Bell is a Data Engineer at Mastodon C, he specialises in high volume streaming systems, BigData solutions and also machine learning applications.

He was section editor for Java Developer’s Journal, contributed to IBM developerWorks on Autonomic Computing and authored the book “Machine Learning: Hands on for Developers and Technical Professionals”.

Leave a Comment or Question

Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?

Join the conversation here (requires login)