Feedzai is a scale up company with one mission: making banking and commerce safe. For that purpose, Feedzai develops methods for fraud prevention that should simultaneously be accurate, scalable and work within low latencies. In this talk will cover Feedzai’s research challenges, focusing on the use of neural networks for fraud prevention.
For a long time, best solutions for fraud detection combined tree-based methods and handmade profile features that represented the user behaviour (ex: number of transactions of a user within the previous week). This approach, however, require handmade engineered features which take time to be computed and which consumes a lot of memory when in production. More recently, with the use of Deep Learning (DL) models with RNNs, we are able to train end-to-end models that encapsulate the feature engineering step in the model itself. These models not only allow for skipping the feature engineering step, but also lead to improvements in typical baseline metrics for fraud prevention, such as transaction recall at 1% false positive rate.