Using Natural Language Processing to Reduce Account Takeovers
According to the LexisNexis 2021 True Cost of Fraud Study1, roughly one-third of financial services companies surveyed responded that account login was the part of the customer journey most susceptible to fraud. When fraud occurs at the account login level, it’s called an account takeover, or ATO. Here at Cash App, we’re always working to improve our algorithms, systems, and...
Recent Link Classification on Temporal Graphs Using Graph Profiler
Cash App customers can interact with each other and with merchants in various ways. They can send payments to one another, gift stocks, purchase and send Bitcoin, or buy products using their Cash App Card. These interactions create a constantly evolving transaction graph. To ensure Cash App remains safe for our customers, we employ advanced machine learning techniques to anticipate...
Real-time Merchant Recommendations on Cash App with Deep Learning
Introduction Cash App thrives on its mission to make customers’ money go further by helping them discover meaningful savings, inspiring them, and connecting them with merchants. This provides a massive opportunity for us to deliver enhanced customer satisfaction and engagement, ultimately leading to better outcomes for the customer, merchants, and Cash App. With that goal, we’re building cutting-edge recommendation models...
Friend Suggestions At Cash App
At Cash App, customers can invite their friends to join the platform and be rewarded for doing so. We built a ML model to populate a “Suggested Friends” section in the Invite Friends screen to help customers quickly find friends to invite without having to scroll through their whole contact book. In this post, we describe how we built it,...
Barbie beats Oppenheimer at the box office - and on Cash App too
Cash App is a vital part of the US economy, providing a crucial tool for people to send and receive Peer to Peer (P2P) payments, obtain a free debit card, and even buy and sell bitcoin and stocks. As data scientists, we’re especially psyched about our robust dataset which gives us an exciting glimpse into how people use Cash App...
Blizzard: A Lightweight, Yet Powerful ML Feature System for Prototyping and v1 Use Cases
Introduction This post is about Blizzard, a machine learning feature system we’ve built at Block’s Cash App for prototyping and early v1 production ML use cases. Blizzard demonstrates how a tremendous amount of value can be obtained by pairing a Python library with a data warehouse. While we already have a robust production-oriented feature system at Cash, that system has...
Improving customer support intent classification with additional language model pretraining
Abstract: Labeled data are a crucial component for many NLP tasks, but are typically time consuming or otherwise costly to collect. Here, we describe how we improved sample efficiency of labeled data for a customer support intent classification task with additional language model pretraining on Cash App support cases.
Cash App goes to Washington
Cash App will be in Washington, D.C. this week, attending the 37th AAAI Conference on Artificial Intelligence. We will be presenting our results on fintech-kMC, a simulation tool we recently built to test and verify machine learning models and workflows using synthetic data. The work will appear as part of the AAAI-23 Bridge Program - AI for Financial Services. We...
Risk at Cash
In the few years since we started, Cash has developed a robust and effective approach to managing customer-facing financial risk. Our approach lets us operate in local markets which have often been ignored by traditional banks, offer products which previously seemed infeasible, and respond to changes rapidly. Fraud is a high-stakes game that takes constant work to stay ahead of...
Model Review
Code Review is an integral part of software development, but many teams across industry don’t have similar processes in place for the development and deployment of Machine Learning (ML) models.