iCubed (Institute, Industry, Innovation) seminars invite DSI Industry Affiliates to give technical talks on work going on in their domain. Join to learn about real-world uses of data science and career opportunities at Industry Affiliate companies.
Guest Speakers
Victoria Martins, Senior Manager, Data Science at Capital One
Daniele Rosa, Senior Manager, Machine Learning Engineer at Capital One
Moderated By: Jessica Rodriguez, Industry Engagement and Outreach Officer, The Data Science Institute
This video recording is private and is available to Columbia University faculty and students only. Please email datascience@columbia.edu for access to the recording.
Talk Information
Nearest Neighbor Search of Merchant Embeddings: A Fraud Use Case
Embeddings are dense vector representations of data that can be used to efficiently compute distances between entities. Embeddings make it trivial to calculate pairwise similarity, cluster, and search for nearest neighbors. As such, these representations can enable many business use cases such as recommendations, search, fraud, and entity resolution. One such use case is the automated detection of merchants at risk of elevated fraud rates. Capital One currently implements transaction defenses against merchants based on historical risk signals. This presentation demonstrates that the nearest neighbor search of the merchant embedding space can be used to identify merchants at risk of ongoing elevated fraud, enabling us to take defensive action earlier.
Victoria Martins is a Senior Manager, Data Science at Capital One where she leads a team building machine learning models for the Enterprise Products & Platforms team. Since joining Capital One in 2019, she has focused on emerging applications of machine learning in digital commerce, merchant intelligence, and fraud defense. Victoria has over ten years of experience in data science and previously led teams at Square and Honey.
Daniele Rosa is a Senior Manager, Machine Learning Engineer at Capital One where he leads a team focused on Graph ML within Applied Research since he joined the company in June 2020. Danele has experience in research and quantitative methods for diverse domains including environmental science, finance, and manufacturing. https://www.linkedin.com/in/drosame
