Topic : DEEP LEARNING APPLICATIONS TO ONLINE PAYMENT FRAUD DETECTION
Abstract: Online data data often poses challenges such as high dimensionality, sparsity, high cardinality and temporal distribution shifts. Fraud detection models using such data require significant tuning to balance model performance and stability so as to address performance play. The talk will cover some applications and use-cases of different deep neural network architectures applied to the problem of online payments fraud detection for addressing such challenges as well as comparative performance benchmarking with other algorithms.
Bio: Distinguished Data Scientist at PayPal
Dates Employed: Apr 2017 – Present Employment
Duration: 5 mos
Location: San Francisco Bay Area
– Payments fraud modeling and research using large scale machine learning algorithms and technologies.
– Applying state-of-the-art algorithms (deep learning using tensorflow and CUDA API’s) and associated ideas in a nuanced manner specific to the payments fraud use cases such as account-take over and credit card fraud