Description: In this informal, but in-depth presentation, Dr. James McCaffrey from Microsoft Research will explain what anomaly detection is and present a complete end-to-end demonstration, implemented from scratch, using the C# language. This talk does not assume you know anything about anomaly detection or neural autoencoders, but does assume you have intermediate or better coding skill. You will leave this presentation with all the information you need to create a practical anomaly detection system for your own data. Topics covered include: anomaly detection fundamentals, data preparation, neural autoencoders, strengths and limitations of autoencoder anomaly detection, and alternatives techniques.

You will learn:

• Understand what anomaly detection is.
• Understand how neural autoencoders work.
• Understand how to implement anomaly detection using raw C#.