Adversarial Machine Learning for Fraud Detection - How Can Organizations Benefit from the Pioneering Work of the NSA and Facebook?
Mounting an effective defense against cyber-attacks requires detection techniques that can evolve as quickly as the attacks themselves. Without the ability to automatically adapt to detect new types of threats, an anti-fraud solution will always be a step behind the fraudsters. Join this session and learn how adversarial machine learning can be effectively used to deter fraud. Dr. Ian Howells, CMO at Argyle Data will walk listeners through how the NSA and Facebook have successfully used adversarial machine learning techniques to deter fraud.
How is technology evolving to analyze multiple and massive streams of data in real time to detect fraudulent activity? The NSA has pioneered data collection techniques at a staggering scale, potentially monitoring all activity for an entire country. Facebook has pioneered adversarial machine-learning fraud detection into an "immune system" that can carry out tens of billions of checks per day to find patterns where the fraudsters are purposefully trying not to create any. We will discuss how the combination and augmentation of these technologies with deep packet inspection enables organizations to deploy them. And we will learn how such a solution could:
- Analyse sparsely populated data sets but that may have millions of distinct features;
- Learn what a negative pattern is where none existed before;
- Perform this analysis in real time and at a massive scale.
This session was recorded during the 2014 Fraud Summit London. Additional recordings include: