Accurate imputation of race and ethnicity (R&E) is crucial for analyzing disparities and informing policy. Methods like Bayesian Improved Surname Geocoding (BISG) are widely used but exhibit limitations, including systematic misclassification biases linked to socioeconomic status. This paper introduces LSTM+Geo, a novel approach enhancing Long Short-Term Memory (LSTM) networks with census tract geolocation information. Using a large voter dataset, we demonstrate that LSTM+Geo (88.7% accuracy) significantly outperforms standalone LSTM (86.4%) and Bayesian methods like BISG (82.9%) and BIFSG (86.8%) in accuracy and F1-score on a held-out validation set. LSTM+Geo reduces the rate at which non-White individuals are misclassified as White (White FPR 19.3%) compared to name-only LSTMs (White FPR 24.6%). While sophisticated ensemble methods incorporating XGBoost achieve the highest overall accuracy (up to 89.4%) and lowest White FPR (17.8%), LSTM+Geo offers strong standalone performance with improved bias characteristics compared to baseline models. Integrating LSTM+Geo into an XGBoost ensemble further boosts accuracy, highlighting its utility as both a standalone model and a component for advanced systems. We give a caution at the end regarding the appropriate use of these methods.

Our technology teams at Zest AI thrive on solving challenging problems that have a big impact on people's daily lives. We build tools to improve consumer lending by meeting the needs of financial institutions and their customers around the world. By leveraging a combination of machine learning and vast amounts of data, we seek to challenge traditional credit scoring models, while expanding the use of our proprietary technology across the fintech industry. Our team is committed to continuously leveraging and adapting the latest advances in AI/ML to benefit those we serve. Machine Learning Engineers at Zest AI build software that enables our customers to quickly create, adopt, and operate machine learning models in production.

Zest AI, a leader in software for credit underwriting, today announced it has been selected as an honoree in Fast Company's inaugural 2021 Next Big Things in Tech Awards. The awards recognize the companies and technologies that promise to redefine their industries and create a positive impact for consumers, businesses and society at large in the next five years. Specifically, Zest AI was recognized for its Fairness Kit, a set of software applications created in response to demand from banks and credit unions looking to reduce bias from consumer lending. Zest's patented machine learning-based solution automatically optimizes credit underwriting models for both accuracy and fairness, creating real options for the first time for lenders that want to close the racial approval rate gap in consumer credit. "We are honored to have our fair lending efforts recognized by Fast Company for this award," said Mike de Vere, CEO of Zest AI. "Traditionally, lenders have had to make trade-offs between accuracy in risk prediction and minimizing disparate impact. Through our technology, these lenders can make more accurate decisions and say yes to more people who might have struggled to get affordable credit. This award further endorses our mission of making fair and transparent credit available to everyone."

Is the US economy headed for an economic slowdown or recession? Forecasting the economy's ups and downs is difficult, but you don't have to batten down the hatches. Attend this webinar to learn how AI can help your lending business weather uncertain market conditions. Tune into hear why forward thinking banking leaders have turned to Zest AI to help them future proof their lending business. Have you created an account yet?

ZestFinance, the leader in artificial intelligence (AI) software for credit, today announced the company's rebrand to Zest AI.